Sys205!1!0 (Hsdpa Protocols)

HSDPA System Architecture and Protocols
HSDPA System Architecture and Protocols FOR TRAINING PURPOSES ONLY - THIS MANUAL WILL NOT BE UPDATED
MOTOROLA LTD.2002
Copyrights The Motorola products described in this document may include copyrighted Motorola computer programs stored in semiconductor memories or other media. Laws in the United States and other countries preserve for Motorola certain exclusive rights for copyright computer programs, including the exclusive right to copy or reproduce in any form the copyright computer program. Accordingly, any copyright Motorola computer programs contained in the Motorola products described in this document may not be copied or reproduced in any manner without the express written permission of Motorola. Furthermore, the purchase of Motorola products shall not be deemed to grant either directly or by implication, estoppel or otherwise, any license under the copyrights, patents or patent applications of Motorola, except for the rights that arise by operation of law in the sale of a product. Restrictions The software described in this document is the property of Motorola. It is furnished under a license agreement and may be used and/or disclosed only in accordance with the terms of the agreement. Software and documentation are copyright materials. Making unauthorized copies is prohibited by law. No part of the software or documentation may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language or computer language, in any form or by any means, without prior written permission of Motorola. Accuracy While reasonable efforts have been made to assure the accuracy of this document, Motorola assumes no liability resulting from any inaccuracies or omissions in this document, or from the use of the information obtained herein. Motorola reserves the right to make changes to any products described herein to improve reliability, function, or design, and reserves the right to revise this document and to make changes from time to time in content hereof with no obligation to notify any person of revisions or changes. Motorola does not assume any liability arising out of the application or use of any product or circuit described herein; neither does it convey license under its patent rights of others. Trademarks Motorola and the Motorola logo are registered trademarks of Motorola Inc. Intelligence Everywhere, M-Cell and Taskfinder are trademarks of Motorola Inc. All other brands and corporate names are trademarks of their respective owners.
MOTOROLA LTD.2002
HSDPA System Architecture and Protocols FOR TRAINING PURPOSES ONLY - THIS MANUAL WILL NOT BE
Contents
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
HSDPA System Architecture and Protocols Chapter 1: HSDPA Overview

Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HSDPA (High Speed Downlink Packet Access) for WCDMA . . . . . . . . . . . . . . . . . Feature Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Compatibility with Release 99 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Demand for Packet Switched Traffic . . . . . . . . . . . . . . . . . . . . . . . . . . . HSDPA Targets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Higher Data Rates for Streaming-, Interactive- and Background Services . . . . . . . . . Consideration of UE Processing Time and Memory Requirements . . . . . . . . . . . . Higher Spectrum Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Small Changes to existing Techniques and Architectures . . . . . . . . . . . . . . . . . Efficient Resource Sharing in Downlink among Users . . . . . . . . . . . . . . . . . . HSDPA Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modulation Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Higher Throughput Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AMC (Adaptive Modulation and Coding) . . . . . . . . . . . . . . . . . . . . . . . . . Hybrid ARQ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transmission and Retransmission Scheduling in NodeB . . . . . . . . . . . . . . . . . QPSK versus 16-QAM Modulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . QPSK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16QAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Maximum Throughput Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UMTS Rel. 99 / Rel. 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HSDPA Rel. 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Important Changes for HSDPA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . New 2 ms Subframe for HSDPA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . New Physical Channels and Transport Channel with HSDPA . . . . . . . . . . . . . . . No Fast Power Control and variable Spreading Factor . . . . . . . . . . . . . . . . . . New UE Capabilities / Categories. . . . . . . . . . . . . . . . . . . . . . . . . . . . . New MAC-hs in NodeB and UE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Impact on NBAP and Frame Protocol Procedure . . . . . . . . . . . . . . . . . . . . . New Channels with HSDPA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transport Channel: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HS-DSCH (High Speed Downlink Shared Channel) . . . . . . . . . . . . . . . . . . Physical Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . High Speed Shared Control Channel (HS-SCCH) . . . . . . . . . . . . . . . . . . . High Speed Physical Downlink Shared Channel (HS-PDSCH) . . . . . . . . . . . . Uplink Dedicated Control Channel Associated with HS-DSCH Transmission (Uplink HS-DPCCH). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HSDPA Basic Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . MAC-hs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . New Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UE Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HSDPA and DPCH Operation HSDPA Setup . . . . . . . . . . . . . . . . . . . . . . . . HSDPA Setup via DPCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HSDPA Transmission and Retransmission . . . . . . . . . . . . . . . . . . . . . . . . Physical Channel Combinations supported with HSDPA on the same FDD Frequency HSDPA High Speed Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H-RNTI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
MOTOROLA LTD.2002
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1-3 1-4 1-4 1-4 1-4 1-6 1-6 1-6 1-6 1-6 1-6 1-8 1-8 1-8 1-8 1-8 1-8 1-10 1-10 1-10 1-12 1-12 1-12 1-14 1-14 1-14 1-14 1-14 1-14 1-14 1-16 1-16 1-16 1-16 1-16 1-16 1-16 1-18 1-18 1-18 1-18 1-20 1-20 1-22 1-22 1-24 1-24
i
Contents
DL-HSPDSCH-Info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HS-SCCH Info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Downlink Scrambling Coding . . . . . . . . . . . . . . . . . . . . . . HS-SCCH Channelization Code info . . . . . . . . . . . . . . . . . . HS-SCCH Codes. . . . . . . . . . . . . . . . . . . . . . . . . . . . Channel Type Switching with HSDPA . . . . . . . . . . . . . . . . . . . . . . CELL_DCH ONLY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CELL_DCH & HSDPA . . . . . . . . . . . . . . . . . . . . . . . . . . . CELL_FACH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HSDPA Control and User Plane Extensions . . . . . . . . . . . . . . . . . . Control Plane Extensions: . . . . . . . . . . . . . . . . . . . . . . . . . New NBAP, RRC & RNSAP IEs . . . . . . . . . . . . . . . . . . . . Radio Resource Sharing . . . . . . . . . . . . . . . . . . . . . . . . Admission / Congestion Control . . . . . . . . . . . . . . . . . . . . User Plane Extensions:. . . . . . . . . . . . . . . . . . . . . . . . . . . PDCP/RLC/MAC-d . . . . . . . . . . . . . . . . . . . . . . . . . . . Frame Protocol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . HSDPA Packet Switched Protocol Stacks . . . . . . . . . . . . . . . . . . . . HSDPA Control Plane . . . . . . . . . . . . . . . . . . . . . . . . . . . Access Stratum Protocols . . . . . . . . . . . . . . . . . . . . . . . Non Access Stratum Protocols . . . . . . . . . . . . . . . . . . . . . MAC-hs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HSDPA User Plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . Configuration without MAC-c/sh . . . . . . . . . . . . . . . . . . . . MAC-d entity UE Side . . . . . . . . . . . . . . . . . . . . . . . . MAC-hs entity UE Side . . . . . . . . . . . . . . . . . . . . . . . . MAC-d entity UTRAN Side . . . . . . . . . . . . . . . . . . . . . . MAC-hs UTRAN Side . . . . . . . . . . . . . . . . . . . . . . . . HSDPA Timing Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . HSDPA Downlink Timing . . . . . . . . . . . . . . . . . . . . . . . . . . DPCH (Dedicated Physical Channel) . . . . . . . . . . . . . . . . . . HS-SCCH (High Speed Shared Control Channel). . . . . . . . . . . . HS-PDSCH (High Speed Physical Downlink Shared Transport Channel) HSDPA Uplink Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5 slots delay of uplink HS-DPCCH . . . . . . . . . . . . . . . . . . Composition of the HS-SCCH . . . . . . . . . . . . . . . . . . . . . . . . . Modulation Type and Number of HS-PDSCHs . . . . . . . . . . . . . . . HARQ Info (Hybrid-ARQ-related Information) . . . . . . . . . . . . . . . . Hybrid ARQ Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . Type I Hybrid ARQ Low UE Complexity . . . . . . . . . . . . . . . . . . Type II Hybrid ARQ High UE Complexity . . . . . . . . . . . . . . . . . Type III Hybrid ARQ medium UE Complexity . . . . . . . . . . . . . . . DCH / HS-DSCH Comparison . . . . . . . . . . . . . . . . . . . . . . . . . Variable Spreading Factor . . . . . . . . . . . . . . . . . . . . . . . . . Fast Power Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AMC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Multicode Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fast Layer 1 ARQ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DPCH / HS-PDSCH Comparison contd. . . . . . . . . . . . . . . . . . . . 16-QAM Modulation / Interleaving / Channel Coding Scheme . . . . . . . . TrCH Multiplexing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Static rate matching - DCH versus HS-DSCH . . . . . . . . . . . . . . . . Transport block concatenation - DCH versus HS-DSCH. . . . . . . . . . . Soft(er) Handover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Advantages and Disadvantages of HSDPA . . . . . . . . . . . . . . . . . . . Advantages of HSDPA . . . . . . . . . . . . . . . . . . . . . . . . . . . Reduced Latency for retransmission . . . . . . . . . . . . . . . . . . Higher DL Throughput and Peak Throughput Rates up to 14 Mbit/s . . . Support of Non-Real Time and Streaming Applications . . . . . . . . . High Spectrum and Code Efficiency due to 16-QAM . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1-26 1-26 1-26 1-26 1-26 1-28 1-28 1-28 1-28 1-30 1-30 1-30 1-30 1-30 1-30 1-30 1-30 1-32 1-32 1-32 1-32 1-32 1-34 1-34 1-34 1-34 1-34 1-34 1-36 1-36 1-36 1-36 1-36 1-38 1-38 1-40 1-40 1-40 1-42 1-42 1-42 1-42 1-44 1-44 1-44 1-44 1-44 1-44 1-46 1-46 1-46 1-46 1-46 1-46 1-48 1-48 1-48 1-48 1-48 1-48
ii
MOTOROLA LTD.2002
Contents
Disadvantages of HSDPA . . . . . . . . . . . . . . . . . . . . . Higher Processing power in UE. . . . . . . . . . . . . . . . . More Memory in UE due to HARQ . . . . . . . . . . . . . . . Advanced Receiver for 16-QAM . . . . . . . . . . . . . . . . Higher Vulnerability using 16-QAM . . . . . . . . . . . . . . . Software and possibly Hardware Upgrade . . . . . . . . . . . No Macro-Diversity for HS-PDSCH, HS-SCCH . . . . . . . . . Power / Interference Increase in uplink and downlink . . . . . . Signalling Overhead caused by HS-SCCH (dl) and DPCCH (ul) . Future Enhancements of HSDPA . . . . . . . . . . . . . . . . . . Beamforming . . . . . . . . . . . . . . . . . . . . . . . . . . Transmit Diversity . . . . . . . . . . . . . . . . . . . . . . . MIMO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preview to HSUPA . . . . . . . . . . . . . . . . . . . . . . . . . Total available Transmission Power. . . . . . . . . . . . . . . Much larger Dynamic of the UL Power Control . . . . . . . . . UL does not suffer from Channelization Code Shortage . . . . . Fast Power Control cannot be abandoned in UL . . . . . . . . HSUPA - Areas of Improvement . . . . . . . . . . . . . . . . . . NodeB controlled scheduling . . . . . . . . . . . . . . . . . . Hybrid ARQ in Uplink . . . . . . . . . . . . . . . . . . . . . . Shorter TTI . . . . . . . . . . . . . . . . . . . . . . . . . . HSUPA - Less Beneficial Areas . . . . . . . . . . . . . . . . . . . Higher Order Modulation . . . . . . . . . . . . . . . . . . . . Fast DCH Setup . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
1-50 1-50 1-50 1-50 1-50 1-50 1-52 1-52 1-52 1-54 1-54 1-54 1-54 1-56 1-56 1-56 1-56 1-56 1-58 1-58 1-58 1-58 1-60 1-60 1-60
Chapter 2: HSDPA Principles

Objectives . . . . . . . . . . . . . . . . . . . . . . . . HS-SCCH Frame Structure . . . . . . . . . . . . . . . . Part 1 . . . . . . . . . . . . . . . . . . . . . . . . Part 2 . . . . . . . . . . . . . . . . . . . . . . . . HS-SCCH Coding Chain . . . . . . . . . . . . . . . . . RV Coding . . . . . . . . . . . . . . . . . . . . . . MUX of HS-SCCH Part 1 . . . . . . . . . . . . . . . MUX of HS-SCCH Part 2 . . . . . . . . . . . . . . . HS-SCCH Channel Coding and Rate Matching . . . . UE specific Masking of Part 1 bits . . . . . . . . . . UE specific Masking for Part 1 . . . . . . . . . . . . . . UE Specific CRC Attachment for Part 1 and Part 2 . . . . HARQ Process ID and New Data Indicator . . . . . . . . HARQ Process ID . . . . . . . . . . . . . . . . . . New Data Indicator . . . . . . . . . . . . . . . . . . 1-Channel Stop and Wait . . . . . . . . . . . . . . . . . HARQ Processes with N-Channel Stop and Wait . . . . . Preconditions: . . . . . . . . . . . . . . . . . . . . HS-SCCH Signaling . . . . . . . . . . . . . . . . . HS-DSCH Transmission and Retransmission . . . . . Multicode Operation in HSDPA . . . . . . . . . . . . . . HS-SCCH-set Decoding . . . . . . . . . . . . . . . HS-DSCH Demodulation . . . . . . . . . . . . . . . HS-SCCH Power Control . . . . . . . . . . . . . . . . . HS-PDSCH Frame Structure . . . . . . . . . . . . . . . HS-PDSCH Channelization Code Tree . . . . . . . . . . Example: . . . . . . . . . . . . . . . . . . . . . . . HSDPA Code- and Time Multiplex Operation . . . . . . . Resource sharing in code as well as well time domain UE procedure for receiving HS-DSCH . . . . . . . . Uplink HS-DPCCH Frame Structure . . . . . . . . . . . HARQ-Ack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3 2-4 2-4 2-4 2-6 2-6 2-6 2-6 2-6 2-6 2-8 2-10 2-12 2-12 2-12 2-14 2-16 2-16 2-16 2-16 2-18 2-18 2-18 2-20 2-22 2-24 2-24 2-26 2-26 2-26 2-28 2-28
MOTOROLA LTD.2002
iii
Contents
Channel Quality Indication . . . . . . . . . . . . . . . . . . . . . . . Uplink HS-DPCCH Code Allocation . . . . . . . . . . . . . . . . . . . . . Multicode Operation in Uplink. . . . . . . . . . . . . . . . . . . . . . Uplink HS-DPCCH Coding Chain . . . . . . . . . . . . . . . . . . . . . . Channel Coding for HS-DPCCH . . . . . . . . . . . . . . . . . . . . Channel Coding for HARQ-Ack . . . . . . . . . . . . . . . . . . . Channel Coding for CQI . . . . . . . . . . . . . . . . . . . . . . Uplink HS-DPCCH Spreading . . . . . . . . . . . . . . . . . . . . . . . Spreading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DPCCH / DPDCH / HS-DPCCH . . . . . . . . . . . . . . . . . . . . Gain Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Uplink Transmission Delay T(0) . . . . . . . . . . . . . . . . . . . . . . . Uplink / Downlink Timing at UE . . . . . . . . . . . . . . . . . . . . . Uplink Fast Power Control Reason for T(0) . . . . . . . . . . . . . . HS-SCCH / HS-DSCH and HS-DPCCH Timing (1) . . . . . . . . . . . . . HS-SCCH / HS-DSCH and HS-DPCCH Timing (2) . . . . . . . . . . . . . Uplink HS-DPCCH . . . . . . . . . . . . . . . . . . . . . . . . . . . T(TX_diff) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Delta Ack / Nack and Ack-Nack Repetition Factor . . . . . . . . . . . . . . DeltaAck . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DeltaNack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ACK-NACK-repetitionFactor . . . . . . . . . . . . . . . . . . . . . . Delta Ack / Nack and CQI Power . . . . . . . . . . . . . . . . . . . . . . HS-DPCCH ACK / NACK and DTX Recognition. . . . . . . . . . . . . . . False Alarm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CQI Description and Task. . . . . . . . . . . . . . . . . . . . . . . . . . TFRC Info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Out of Range Signaling . . . . . . . . . . . . . . . . . . . . . . . . . Power Reduction Factor 1 . . . . . . . . . . . . . . . . . . . . . . . Measurement Power Offset 2. . . . . . . . . . . . . . . . . . . . . . CQI Reporting Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . HS-DPCCH Measurement Feed Back - CQI . . . . . . . . . . . . . . . . Measurement Power Offset 3 . . . . . . . . . . . . . . . . . . . . . . CQI Feedback cycle, k . . . . . . . . . . . . . . . . . . . . . . . . . CQI-RepetitionFactor . . . . . . . . . . . . . . . . . . . . . . . . . . DeltaCQI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CQI Reporting Principle . . . . . . . . . . . . . . . . . . . . . . . . . . UE measures the downlink channel quality . . . . . . . . . . . . . . . CQI Algorithm in UE . . . . . . . . . . . . . . . . . . . . . . . . . . NodeB TFRC decision . . . . . . . . . . . . . . . . . . . . . . . . . CQI Mapping Table UE Category . . . . . . . . . . . . . . . . . . . . . UE Category . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CQI value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transport Block Size . . . . . . . . . . . . . . . . . . . . . . . . . . Reference Power Adjustment. . . . . . . . . . . . . . . . . . . . . . N(IR) and X(RV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . Example for a CQI Report . . . . . . . . . . . . . . . . . . . . . . . . . CQI and Ack/Nack Considerations . . . . . . . . . . . . . . . . . . . . . HS-DPCCH Power Control during Soft Handover . . . . . . . . . . . . CQI Repetition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Intra NodeB Softer HO . . . . . . . . . . . . . . . . . . . . . . . . . CQI Reporting Feedback Cycle and Repetition . . . . . . . . . . . . . . Ack/Nack Repetition Factor . . . . . . . . . . . . . . . . . . . . . . . Trade off between Ack/Nack/CQI Repetition and long Feedback Cycles . HSDPA during Compressed Mode Operation . . . . . . . . . . . . . . . . Effects of Using 16-QAM . . . . . . . . . . . . . . . . . . . . . . . . . . Higher Throughput Rates. . . . . . . . . . . . . . . . . . . . . . . . Increased Spectrum Efficiency . . . . . . . . . . . . . . . . . . . . . Higher Interference Vulnerability . . . . . . . . . . . . . . . . . . . . Smaller Decision Space . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2-28 2-30 2-30 2-32 2-32 2-32 2-32 2-34 2-34 2-34 2-34 2-36 2-36 2-36 2-38 2-40 2-40 2-40 2-42 2-42 2-42 2-42 2-44 2-46 2-46 2-48 2-48 2-48 2-48 2-48 2-48 2-50 2-50 2-50 2-50 2-50 2-52 2-52 2-54 2-56 2-58 2-58 2-58 2-58 2-58 2-58 2-60 2-62 2-62 2-62 2-62 2-64 2-64 2-64 2-66 2-68 2-68 2-68 2-68 2-68
iv
MOTOROLA LTD.2002
Contents
16-QAM Constellation Rearrangement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-QAM Constellation Rearrangement for b = 2 and b = 3 . . . . . . . . . . . . . . . . . . . . Bit Re-arrangement b = 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bit Re-arrangement b = 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Power Sharing between HSDPA and DCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . Node B Transmit Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Maximum Transmission Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UE Transmit Power. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adaptive Modulation and Coding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Main Benefits of AMC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Availability of higher throughput and therefore increased average throughput per cell. . . . Reduced interference variation due to modulation and coding based link adaptation instead of variations in transmit power. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Utilization of short term fading in the way that a user might be always served in a constructive fade. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Boosted effectiveness in combination with time and code multiplexing scheduling techniques.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adaptive Modulation and Coding - Principle. . . . . . . . . . . . . . . . . . . . . . . . . . . . AMC Scheduling Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Layer 1 Processing Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Physical Layer Processing Chain of AMC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods of selecting the suitable MCS . . . . . . . . . . . . . . . . . . . . . . . . . . . . AMC Processing Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AMC Liabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UEs measurement inaccuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Delay in CQI Reporting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Limited Granularity for Data Rate Selection . . . . . . . . . . . . . . . . . . . . . . . . . . NodeB Processing and Scheduling Delay. . . . . . . . . . . . . . . . . . . . . . . . . . . HSDPA HARQ Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chase Combining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Partial IR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Full IR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HARQ Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chase Combining Performance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Operation of Partial IR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Operation of Full IR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scheduling Strategies Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AMC Scheduling Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The predicted or estimated channel quality . . . . . . . . . . . . . . . . . . . . . . . . The current load of the cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The traffic priority class (e.g. streaming class or background class) . . . . . . . . . . . . QoS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparison of Basic Packet Scheduler Methods . . . . . . . . . . . . . . . . . . . . . . . . . Scheduling Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Serve Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Allocation Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . FT (Fair Throughput) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . P-FR (Proportional Fair Resources) . . . . . . . . . . . . . . . . . . . . . . . . . . . M-C/I (Maximum C/I) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Packet Scheduler in NodeB. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fairness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Opportunistic Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Minimum Throughput Guarantees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effect of TCP/IP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Influence of Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Packet Scheduling Strategies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fair Throughput Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fair Time Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Long Term C/I Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Proportional Fair Resource Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2-70 2-72 2-72 2-72 2-74 2-74 2-74 2-74 2-76 2-76 2-76 2-76 2-76 2-76 2-78 2-78 2-78 2-80 2-80 2-80 2-82 2-82 2-82 2-82 2-82 2-84 2-84 2-84 2-84 2-86 2-86 2-88 2-90 2-92 2-92 2-92 2-92 2-92 2-92 2-94 2-94 2-94 2-94 2-94 2-94 2-94 2-96 2-96 2-96 2-96 2-96 2-96 2-98 2-98 2-98 2-98 2-98
v
MOTOROLA LTD.2002
Contents
Multi-User Selection Diversity Dynamic Scheduling . . . . . . . . . . . . . . . . Dynamic Scheduling according to best C/I . . . . . . . . . . . . . . . . . . . Scheduler Resources include: . . . . . . . . . . . . . . . . . . . . . . . . . MAC-hs Scheduler should not allocate any HSDPA Resources to a User which:
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
2-100 2-100 2-100 2-100
Chapter 3: Forward and Backward Error Correction

Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Turbo Coding. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Turbo Coder Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RSC Encoder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interleaver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Turbo Coder Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Trellis termination for Turbo coder . . . . . . . . . . . . . . . . . . . . . . . . . . Turbo Code internal Interleaver . . . . . . . . . . . . . . . . . . . . . . . . . . . . Iterative Decoding Principle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . TFRI - Transport Block Size Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . QPSK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-QAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HARQ Retransmission Comparison with Legacy Releases . . . . . . . . . . . . . . . DCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HS-DSCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Redundancy Version and Constellation Version . . . . . . . . . . . . . . . . . . . . . . HSDPA Category and Reference Combinations. . . . . . . . . . . . . . . . . . . . . . HS-DSCH Category . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Maximum number of HS-DSCH multi codes . . . . . . . . . . . . . . . . . . . . . Minimum inter-TTI interval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Maximum Number of HS-DSCH Transport Channel Bits received within HS-DSCH TTI Total Number of Soft Channel Bits over all HARQ Processes . . . . . . . . . . . . . Maximum Number of RLC-AM Entities . . . . . . . . . . . . . . . . . . . . . . . . Total RLC-AM and MAC-hs Buffer Size [KBytes] . . . . . . . . . . . . . . . . . . . Number of HARQ Processes versus IR Memory . . . . . . . . . . . . . . . . . . . . . Chase Combining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IR Retransmission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HARQ Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Memory Partitioning is set to Implicit . . . . . . . . . . . . . . . . . . . . . . . . . Memory Partitioning is set to Explicit. . . . . . . . . . . . . . . . . . . . . . . . . HARQ Principle in HSDPA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . NodeB Side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UE Side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HS-DSCH HARQ Functionality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Turbo Coding. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bit Separation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1st RM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2nd RM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bit Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HARQ Parameters for Retransmission . . . . . . . . . . . . . . . . . . . . . . . . . . Transmission Parameters = Retransmission Parameters . . . . . . . . . . . . . . . Transmission Parameters <> Retransmission Parameters . . . . . . . . . . . . . . Rate Matching Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . First Rate Matching. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Second Rate Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N(CH) = N(data) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bit Rearrangement and Redundancy Version Selection . . . . . . . . . . . . . . First RM Stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Puncturing Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Second Rate Matching Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RM Pattern Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bit Repetition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3 3-4 3-4 3-4 3-4 3-6 3-6 3-6 3-8 3-10 3-10 3-12 3-14 3-14 3-14 3-16 3-18 3-18 3-18 3-18 3-18 3-18 3-18 3-18 3-20 3-20 3-20 3-22 3-22 3-22 3-24 3-24 3-24 3-26 3-26 3-26 3-26 3-26 3-26 3-26 3-28 3-28 3-28 3-30 3-30 3-30 3-30 3-30 3-32 3-32 3-34 3-36 3-36
vi
MOTOROLA LTD.2002
Contents
Example for Initial Transmission Self-decodable with 1st RM . Example for Initial Transmission-Self-decodable with 1st RM Example for Retransmission Non-self-decodable . . . . . . . Full IR . . . . . . . . . . . . . . . . . . . . . . . . . . . Partial IR . . . . . . . . . . . . . . . . . . . . . . . . . . Example for Chase Combining . . . . . . . . . . . . . . . . . Overview over transmission capacity considerations . . . . 2nd RM Parameter Calculation Self-Decodable . . . . . . . . Turbo Coding with Transparent 1st RM stage . . . . . . . . X(i), e(plus) and e(minus) Calculation. . . . . . . . . . . . Initial Setting of e and Redundancy Version r . . . . . . . . 2nd RM Bit Position Calculation Self-Decodable . . . . . . . Systematic RM . . . . . . . . . . . . . . . . . . . . . . . Parity_1 RM . . . . . . . . . . . . . . . . . . . . . . . . Parity_2 RM . . . . . . . . . . . . . . . . . . . . . . . . 2nd RM Self-Decodable Transmission followed by Full IR (1) . 2nd RM Self-Decodable Transmission followed by Full IR (2) . 2nd RM Parameter Calculation Non-Self-Decodable . . . . . 2nd RM Bit Position Calculation Non-Self-Decodable . . . . . Rate Matching Exercise . . . . . . . . . . . . . . . . . . . . Comparison between Full and Partial IR . . . . . . . . . . . . R = 3/4 . . . . . . . . . . . . . . . . . . . . . . . . . . . R = 1/2 . . . . . . . . . . . . . . . . . . . . . . . . . . . HARQ Performance Enhancement Techniques . . . . . . . . . HS-PDSCH Interleaving . . . . . . . . . . . . . . . . . . . . DTX Indication Bits . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
3-38 3-38 3-40 3-40 3-40 3-42 3-42 3-44 3-44 3-44 3-44 3-46 3-46 3-46 3-46 3-48 3-50 3-52 3-54 3-56 3-58 3-58 3-58 3-60 3-62 3-62
Chapter 4: HSDPA Protocol Enhancements and Extensions

Objectives . . . . . . . . . . . . . . . . . . . . . . . . HSDPA Architecture Evolution from Rel 99 / Rel 4 . . . Protocol Structure . . . . . . . . . . . . . . . . . . HS-DSCH Characteristics . . . . . . . . . . . . . . NBAP - HSDPA Message and IE Overview . . . . . . . . HS-DSCH Resources Information . . . . . . . . . . HS-DSCHs MAC-d Flow To Rearrange. . . . . . . . HS-DSCH FDD Information. . . . . . . . . . . . . . HS-DSCH MAC-d Flows To Add . . . . . . . . . . . HS-DSCH Information To Modify . . . . . . . . . . . HS-DSCH MAC-d Flows To Delete . . . . . . . . . . NBAP - HSDPA Common Measurement Report. . . . . . HS-DSCH Required Power Value Information. . . . . Scheduling Priority Indicator . . . . . . . . . . . HS-DSCH Required Power Value. . . . . . . . . HS-DSCH Required Power Per UE Information . . . . HS-DSCH Required Power Per UE Weight . . . . HS-DSCH Provided Bit Rate Value Information . . HS-DSCH Provided Bit Rate Value . . . . . . . . NBAP HSDPA Radio Link Parameter Update Indication . NBAP - HS-DSCH Related IEs (1) . . . . . . . . . . . . Allocation / Retention Priority . . . . . . . . . . . . . ARP Value Meaning . . . . . . . . . . . . . . . Priority Flow Information . . . . . . . . . . . . . . . Scheduling Priority Indicator . . . . . . . . . . . Discard Timer . . . . . . . . . . . . . . . . . . MAC-hs Guaranteed Bit Rate. . . . . . . . . . . NBAP - HS-DSCH Related IEs (2) . . . . . . . . . . . . MAC-d PDU Size. . . . . . . . . . . . . . . . . . . RLC Mode . . . . . . . . . . . . . . . . . . . . UE Capabilities Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3 4-4 4-4 4-4 4-6 4-6 4-6 4-6 4-6 4-6 4-6 4-8 4-8 4-8 4-8 4-8 4-8 4-8 4-8 4-10 4-12 4-12 4-12 4-12 4-12 4-12 4-12 4-14 4-14 4-14 4-14
MOTOROLA LTD.2002
vii
Contents
HS-SCCH Power Offset . . . . . . . . . . . . NBAP HS-DSCH Info Response . . . . . . . . . . . . HS-DSCH Initial Capacity Allocation . . . . . . . . HS-SCCH Specific Information Response . . . . . HARQ Memory Partitioning Information . . . . . . . HS-PDSCH and HS-SCCH Power and Code Information HSDPA Cell Configuration - AUDIT . . . . . . . . . . . HSDPA Data Transfer Procedure (1) . . . . . . . . . . HSDPA Data Transfer Procedure (2) . . . . . . . . . . RRC Protocol Enhancement New IEs with HSDPA . . MP Mandatory Present . . . . . . . . . . . . . . MD Mandatory Default . . . . . . . . . . . . . . CV Conditional Value . . . . . . . . . . . . . . . OP - Optional. . . . . . . . . . . . . . . . . . . . Added or Reconfigured MAC-d Flow . . . . . . . . . . MAC-d Flow Identity . . . . . . . . . . . . . . . . MAC-hs Queue Id . . . . . . . . . . . . . . . . . T1 Reorder Release Timer . . . . . . . . . . . . . MAC-d PDU Size Info . . . . . . . . . . . . . . . Uplink DPCH Power Control Info and Others . . . . . . Uplink DPCH Power Control Info . . . . . . . . . . MAC-hs Reset . . . . . . . . . . . . . . . . . . . Serving HS-DSCH Radio Link Indicator . . . . . . . MAC-hs Protocol . . . . . . . . . . . . . . . . . . . . MAC-hs - co-incident CRNC and SRNC . . . . . . HS-DSCH RNTI . . . . . . . . . . . . . . . . . . MAC-hs Tasks . . . . . . . . . . . . . . . . . . . . . Iub Flow Control . . . . . . . . . . . . . . . . . . Packet Buffering and Scheduling . . . . . . . . . . TFRC Selection . . . . . . . . . . . . . . . . . . AMC . . . . . . . . . . . . . . . . . . . . . . . . HARQ . . . . . . . . . . . . . . . . . . . . . . . HS-DSCH MAC PDU . . . . . . . . . . . . . . . . . . General . . . . . . . . . . . . . . . . . . . . . . MAC-hs SDU. . . . . . . . . . . . . . . . . . . . MAC-hs Header of MAC-hs PDU . . . . . . . . . . MAC-hs Entity UTRAN Side . . . . . . . . . . . . . . Scheduling/Priority Handling . . . . . . . . . . . . HARQ . . . . . . . . . . . . . . . . . . . . . . . TFRC Selection . . . . . . . . . . . . . . . . . . MAC-d / Flow Control. . . . . . . . . . . . . . . . MAC-hs Entity UE Side . . . . . . . . . . . . . . . . . HARQ . . . . . . . . . . . . . . . . . . . . . . . Reordering Queue distribution . . . . . . . . . . . Reordering . . . . . . . . . . . . . . . . . . . . . Disassembly . . . . . . . . . . . . . . . . . . . . MAC-d . . . . . . . . . . . . . . . . . . . . . . . HS-DSCH MAC Architecture UTRAN Side . . . . . . Overall Architecture . . . . . . . . . . . . . . . . MAC-c/sh (optional) . . . . . . . . . . . . . . . . MAC-d Flow in UTRAN . . . . . . . . . . . . . . . MAC-d flow: 0 . . . . . . . . . . . . . . . . . MAC-d flow: 1 . . . . . . . . . . . . . . . . . MAC-d flow: 2 . . . . . . . . . . . . . . . . . HS-DSCH Architecture UE Side . . . . . . . . . . . Overall Architecture . . . . . . . . . . . . . . . . MAC-d Flow in UE . . . . . . . . . . . . . . . . . Reordering / Disassembly . . . . . . . . . . . MAC-d flow: 0, 1, 2. . . . . . . . . . . . . . . HS-DSCH Scheduler Functions. . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4-14 4-16 4-16 4-16 4-16 4-18 4-20 4-22 4-24 4-26 4-26 4-26 4-26 4-26 4-28 4-28 4-28 4-28 4-28 4-30 4-30 4-30 4-30 4-32 4-32 4-32 4-34 4-34 4-34 4-34 4-34 4-34 4-36 4-36 4-36 4-36 4-38 4-38 4-38 4-38 4-38 4-40 4-40 4-40 4-40 4-40 4-40 4-42 4-42 4-42 4-42 4-42 4-42 4-42 4-44 4-44 4-44 4-44 4-44 4-46
viii
MOTOROLA LTD.2002
Contents
UE Scheduling . . . . . . . . . . . . . . . . . . . . . . Serving Priority Queues . . . . . . . . . . . . . . . . . Updating MAC-hs Header . . . . . . . . . . . . . . . . Scheduling new Transmission and retransmission . . . . Determining RV . . . . . . . . . . . . . . . . . . . . . NodeB HARQ . . . . . . . . . . . . . . . . . . . . . . . . HARQ Entity . . . . . . . . . . . . . . . . . . . . . . . HARQ Process . . . . . . . . . . . . . . . . . . . . . . Error Handling in HARQ . . . . . . . . . . . . . . . . . HS-DSCH Bit Rate Measurement. . . . . . . . . . . . . . . UE MAC-hs Operation . . . . . . . . . . . . . . . . . . . . UE HARQ . . . . . . . . . . . . . . . . . . . . . . . . HARQ Entity . . . . . . . . . . . . . . . . . . . . . HARQ Process . . . . . . . . . . . . . . . . . . . . Transmitter and Receiver Stall Avoidance. . . . . . . . . . . Timer-Based Scheme . . . . . . . . . . . . . . . . . . Window Bases Scheme . . . . . . . . . . . . . . . . . HARQ Activity Scheme . . . . . . . . . . . . . . . . . . UE Reordering Entity . . . . . . . . . . . . . . . . . . . . . Parameters. . . . . . . . . . . . . . . . . . . . . . . . State Variables . . . . . . . . . . . . . . . . . . . . . . Timer T1 . . . . . . . . . . . . . . . . . . . . . . . . . Receiver Window. . . . . . . . . . . . . . . . . . . . . Timer and / or Window Based Stall Avoidance . . . . . . . . Window Based Scheme . . . . . . . . . . . . . . . . . Timer Bases Scheme. . . . . . . . . . . . . . . . . . . UE MAC-hs Other Functions . . . . . . . . . . . . . . . . . Disassemble Entity . . . . . . . . . . . . . . . . . . . . MAC-hs Reset . . . . . . . . . . . . . . . . . . . . . . HS-PDSCH and DL TrCH Information. . . . . . . . . . . . . Activation Time . . . . . . . . . . . . . . . . . . . . . . Downlink Transport Channel Type . . . . . . . . . . . . HS-DSCH Data Frame on Iub/Iur (HS-DSCH FP) . . . . . . . CmCH-PI (Common Transport Channel Priority Indicator). MAC-d PDU Length . . . . . . . . . . . . . . . . . . . NumOfPDU . . . . . . . . . . . . . . . . . . . . . . . User Buffer Size . . . . . . . . . . . . . . . . . . . . . MAC-d PDU . . . . . . . . . . . . . . . . . . . . . . . Control Frame Format for HS-DSCH FP . . . . . . . . . . . HS-DSCH Capacity Request Control Frame . . . . . . . HS-DSCH Capacity Allocation Control Frame . . . . . . . . . Maximum MAC-d PDU Length . . . . . . . . . . . . . . HS-DSCH Credits . . . . . . . . . . . . . . . . . . . . HS-DSCH Interval . . . . . . . . . . . . . . . . . . . . HS-DSCH Repetition Period . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4-46 4-46 4-46 4-46 4-46 4-48 4-48 4-48 4-48 4-50 4-52 4-52 4-52 4-52 4-54 4-54 4-54 4-54 4-56 4-56 4-56 4-56 4-56 4-58 4-58 4-58 4-60 4-60 4-60 4-62 4-62 4-62 4-64 4-64 4-64 4-64 4-64 4-64 4-66 4-66 4-68 4-68 4-68 4-68 4-68
Chapter 5: HSDPA Mobility

Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction to HSDPA Mobility Procedures . . . . . . . . . . Serving HS-DSCH Cell Change . . . . . . . . . . . . . Best Serving HS-DSCH Cell Measurement . . . . . . . . . . Change of best Cell . . . . . . . . . . . . . . . . . . . HS-DSCH Handover . . . . . . . . . . . . . . . . . . . Intra NodeB Synchronized HS-DSCH Cell Change . . . . . . MAC-hs Preservation. . . . . . . . . . . . . . . . . . . Intra NodeB Synchronized HS-DSCH Cell Change Procedure Inter NodeB Synchronized Serving HS-DSCH Cell Change . . Inter NodeB HS-DSCH Cell Change Hard Handover(1) . . . Inter NodeB HS-DSCH Cell Change Hard Handover (2). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3 5-4 5-4 5-6 5-6 5-6 5-8 5-8 5-10 5-12 5-14 5-16
MOTOROLA LTD.2002
ix
Contents
This page intentionally left blank.
MOTOROLA LTD.2002
About This Manual
Version 1 Rev 0

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
HSDPA Overview
Version 1 Rev 0
Chapter 1 HSDPA Overview
MOTOROLA LTD.2002
1-1
Version 1 Rev 0
HSDPA Overview
1-2
MOTOROLA LTD.2002
Objectives
Version 1 Rev 0
Objectives
On completion of this chapter the Student will be able to: Describe the important changes and characteristics of HSDPA State the new channels of HSDPA and how they operate in principle Describe the extended UTRAN protocol stack with HSDPA State the advantages and disadvantages of HSDPA and future enhancements Describe the concept of HSUPA
MOTOROLA LTD.2002
1-3
Version 1 Rev 0
HSDPA (High Speed Downlink Packet Access) for WCDMA

HSDPA considers the trend that the volume of IP-based traffic has already exceeded that for circuit-switched traffic in most fixed networks. The same change can be anticipated in mobile networks because of new IP-based mobile services becoming available and are used by increasing number of people in their daily communication. Current estimates show that in advanced mobile communication markets, packet-switched traffic will overtake circuit-switched traffic in the near future. Delivery of digital content over mobile networks will generate additional traffic and revenue.
Feature Study
The HSDPA feature in 3GPP Release 5 is the result of a study carried out in the Release 4 time frame. This study considered a number of techniques in order to provide instantaneous high speed data in the downlink. Some of the considerations and goals taken into account in the evaluation of the different techniques were: To focus on the streaming, interactive and background services: services which require a constant and/high throughput or low error rate. To prioritise urban environments and then indoor deploymens (but not limited to these environments and supporting full mobility). To enable compatibility with advanced antenna and receiver techniques: transmit and receive diversity methods are used and might be enhanced To take into account User Equipment processing time and memory requirements: UEs limitations are taken into account by the network To minimize changes on existing techniques and architectures: modest changes to NodeB hardware and UTRAN software
Compatibility with Release 99

HSDPA is designed to co-exist on the same carrier as the current Release 99 WCDMA services, enabling a smooth and cost-efficient introduction of HSDPA into existing WCDMA networks.
Demand for Packet Switched Traffic

The increasing demand for capacity in order to provide high data rate multimedia services in wireless environments necessitates enhanced radio transmission techniques and network protocol functionality. Such techniques have to be added to already existing mobile cellular networks. For 3rd generation UMTS networks based on WCDMA, the HSDPA is being introduced to meet this demand and improve spectral efficiency by higher order modulation using 16-QAM. Note: HSDPA achieves gross data rates in downlink up to 14 Mbit/s under ideal conditions. The reverse link (uplink) may remain on 64 kbit/s until at a later stage new features are introduced with HSUPA (High Speed Uplink Packet Access). However, HSUPA will require new mobile terminals and PC-cards to operate with. Except where otherwise indicated, the description of this course only applies to the FDD mode of UMTS.
1-4
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-5
Version 1 Rev 0
HSDPA Targets
HSDPA Targets
Higher Data Rates for Streaming-, Interactive- and Background Services
HSDPA is a feature based on a downlink shared channel that allows user net-data rates of up to 10 Mbit/s. It is designed to support services that require instantaneous high rates in the downlink and lower rates on the uplink. This feature also decreases the level of retransmissions (at the radio link and hence higher layers), in turn allowing the reduction of delivery time. Examples of end-user services targeted by HSDPA are internet browsing and video on demand.
Consideration of UE Processing Time and Memory Requirements

HSDPA takes UE limitations like available physical memory for transmission and especially for retransmission into account. Also the physical channel processing capability is considered. (Examples: Minimum inter-TTI interval, transport channel bits per TTI)
Higher Spectrum Efficiency

With 16-QAM applied in downlink, throughput rates can be doubled compared to QPSK which is used for Rel. 99 and Rel. 4 physical channels. The amount of bits/Hz is increased with 16-QAM as one modulation symbol corresponds to 4 chips whereas in QPSK one modulation symbol represents 2 chips. Even when HSDPA is uses QPSK modulation the spectrum efficiency increases as HSDPA exploits good C/I conditions. This is achieved by reducing the protection (increasing the code rate) and thus having more capacity for the application data.
Small Changes to existing Techniques and Architectures

HSDPA minimizes the necessary upgrades and changes in UTRAN and UE. Nevertheless some protocol additions are necessary in NodeB and UE as well some enhancements of existing procedures and protocols. They are mentioned in the following pages and are explained in full detail in chapter 4.
Efficient Resource Sharing in Downlink among Users

HSDPA introduces a new transport channel type that makes efficient use of valuable radio frequency resources. Beside this, it takes into account the bursty nature of packet switched data by sharing the channelization codes, transmission power and infrastructure hardware among users.
1-6
MOTOROLA LTD.2002
HSDPA Targets
Version 1 Rev 0
HSDPA Targets
MOTOROLA LTD.2002
1-7
Version 1 Rev 0
HSDPA Characteristics
High Speed Downlink Packet Access comes with certain characteristics which distinguishes it clearly from Rel. 99 UTRAN.
Modulation Types
QPSK is already known from Rel. 99 UTRAN. Besides QPSK, HSDPA incorporates the 16-QAM modulation to increase the peak data rates for users served under favorable radio conditions. Support for QPSK is mandatory, though the support for 16-QAM is optional for the network and UE. 16-QAM (Quadrature Amplitude Modulation) was newly introduced in Rel .5 with HSDPA. It is a so called higher order modulation which basically doubles the data rate in good radio conditions. Thus it increases the spectrum efficiency of WCDMA.
Higher Throughput Rates

HDSPA supports peak throughput rates far beyond 2 Mbit/s when radio conditions are suitable and therefore it satisfies the demand for instantaneous high throughput of packet switched services e.g. streaming or interactive traffic class. Theoretically, under optimum condition (Code Rate of 1:1) the following maximum throughput can be achieved: (with 16-QAM (Quadrature Amplitude Modulation) and 15 channelization codes simultaneously used) 960kbit/s x 15 = 14.4 Mbit/s
AMC (Adaptive Modulation and Coding)

AMC is a key feature of HSDPA allowing adjustment of modulation between QPSK and 16-QAM according to radio conditions and retransmission ratio. In addition a variable code rate is used to flexibly adapt the data rate to the physical channel capacity depending on the UEs downlink C/I..
Hybrid ARQ
HARQ functionality combines retransmission with the original transmissions. There a two different ways for HARQ to operate. Either identical retransmission of the data block are sent or retransmission are not identical and differ in data and parity bits compared to the original transmission. The first method is known as chase combining and, the latter as incremental redundancy . HARQ operates on an N-channel Stop and Wait principle.
Transmission and Retransmission Scheduling in NodeB

All Rel. 99 transport channels are terminated at the RNC, except BCH; hence the retransmission procedure for packet data is located in the serving RNC. In order to maximize throughput and reduce delays when retransmitting, additional intelligence is put into the NodeB. In this way, retransmission is controlled by the NodeB, leading to faster retransmission and therefore shorter delay for packet data services. A scheduler in NodeB evaluates for different users what the radio channel conditions are, how much data is pending for each user, how much time has passed since a particular user was last served, for which user retransmission are pending etc. From this input data the scheduler in NodeB may derive a decision how to assign resources to certain users.
1-8
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-9
Version 1 Rev 0
QPSK versus 16-QAM Modulation

The figure illustrates the I/Q Plane for QPSK and 16-QAM modulation technique.
QPSK
Each symbol corresponds to 2 consecutive input bits. The four symbols are represented by different phase shifts in the I/Q plane.
16QAM
Each symbol corresponds to four consecutive input bits. Thus the data rate can be doubled with 16-QAM compared to QPSK. The 16 symbols are represented in the I/Q plane by different phase shifts and amplitudes. In 16-QAM modulation the symbol value is determined by phase and amplitude. Compared to that, in QPSK the phase is only modulated and variation in amplitude have only minor influence on the decision space in the I/Q diagram. However with 16-QAM the decision space is heavily influenced by amplitude variations, thus higher constraints are put on the transmitter linearity. Note, a more accurate phase estimate is necessary with 16-QAM since constellation points have smaller differences in phase domain compared to QPSK. Note: The number of constellation points in the I/Q-diagram can be calculated with 2m, where m represents the number of bits or chips per modulation symbol. QPSK modulation has four constellation points in the I/Q-diagram: 2^m = 4 m = 2. 16-QAM modulation has 16 constellation points in the I/Q-diagram: 2^m = 16 m = 4
1-10
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-11
Version 1 Rev 0
Maximum Throughput Rates

The formulas opposite consider the physical maximum throughput rates available in FDD mode of WCDMA according to releases and modulation. In Rel. 99 and Rel. 4 only QPSK is used whereas Rel. 5 allows also to user higher order modulation scheme 16-QAM. The standard chip rate is 3.84 Mchips/s across all releases. The slot duration is in all releases 0.67ms corresponding to 2560 chips. Note that the physical maximum chip rate achievable with 3.84 Mchips/s and QPSK modulation is: 2 chips/symbol x 3.84 Mchips/s = 7.68 Mchips/s
UMTS Rel. 99 / Rel. 4

Downlink: In the downlink, the DPDCH and DPCCH are time multiplexed onto I and Q plane so the DPDCH data rate also depends on the DPCCH data rate. The physical maximum bit rate (ignoring losses due to DPCCH) using spreading factor 4 is 5.76 Mbit/s. The downlink slot format # 16 allows for 1248 DPDCH bits/slot and per physical channel. The maximum DPDCH data rate considering 3 spreading codes @ sf4 is: 1248 bits/slot x 15 slots x (3 OVSFs) = 5.616 Mbit/s Uplink: In the uplink each channel DPDCH and DPCCH is assigned an orthogonal channelization code. As both physical channels are I/Q multiplexed, (i.e. separated onto I and Q phases), the maximum physical data rate has to be calculated with 1 bit/symbol. The uplink slot format # 6 allows for 640 bits/slot. The maximum DPDCH data rate considering 6 spreading codes @ sf4 is therefore: 5.76 Mbit/s.
HSDPA Rel. 5
In HSDPA the spreading factor for the user plane is fixed to 16. Thus up to 15 physical channels can be allocated at maximum per UE. QPSK: The downlink slot format # 0 allows for 320 bits/slot and per physical channel. This results in 960 bits/ 2 ms subframe. The maximum physical data rate considering 15 spreading codes @ sf16 is: 7.2 Mbit/s. 16-QAM The downlink slot format #1 allows for 640 bits/slot and per physical channel. This results in 1920 bits/ 2 ms subframe. The maximum physical data rate considering 15 spreading codes @ sf16 is: 14.4 Mbit/s.
1-12
MOTOROLA LTD.2002
Version 1 Rev 0
UMTS Rel. 99 / Rel 4

Downlink
2 bits/symbol x [3.84 Mcps / (4 chips/symbol)] x (3 OVSF's) = 5.76 Mbits/S
Uplink
HSDPA Rel. 5
QPSK
16-QAM
MOTOROLA LTD.2002
1-13
Version 1 Rev 0
Important Changes for HSDPA

HSDPA involves significant changes in the UTRAN providing a high flexibility to react to changing air-interface conditions or variable user QoS.
New 2 ms Subframe for HSDPA

The TTI (Transmission Time Interval) in HSDPA has been reduced to 2 ms in order to be faster in retransmitting erroneous data blocks compared to the minimum TTI of 10 ms in UTRA-FDD. Another advantage of the shorter TTI in HSDPA is that NodeB can adapt literally every data block to fast changing radio conditions by the means of AMC. Thus it is possible to counteract fading on the air-interface by adjusting modulation and coding every 2 ms.
New Physical Channels and Transport Channel with HSDPA

New channels are introduced for HSDPA: HS-PDSCH, HS-SCCH, HS-DPCCH and HS-DSCH.
No Fast Power Control and variable Spreading Factor

With HSDPA, two of the most fundamental features of WCDMA, fast power control and variable spreading factor are disabled and replaced by AMC (Adaptive Modulation and Coding). Note: AMC uses multicode operation (the UE can use more than one channelization code in parallel) in order to increase the data rate for a certain user and adapts the code rate to the air-interface quality. By these means AMC is able to improve the user throughput or at least keep it constant even the downlink channel quality deteriorates between subsequent transmissions.
New UE Capabilities / Categories

The HSDPA feature is optional for both UE and network in Rel. 5. The UE indicates its HSDPA support and its HS-DSCH physical layer category within the radio access capability parameter.. The physical layer category defines among other parameters the maximum number of channelization codes the UE supports in parallel for multicode operation. A UE may support up to 5, 10 or 15 channelization codes in parallel.
New MAC-hs in NodeB and UE

The implementation of Medium Access Control (MAC-hs) in NodeB and UE is a pre-requisite for allowing the NodeB to schedule transmissions and retransmission, to maintain the HSDPA specific channels and to operate with AMC and Hybrid ARQ.
Impact on NBAP and Frame Protocol Procedure

NBAP procedures need to support HSDPA capability and HSDPA related parameters. The increased bandwidth needs to be supported by the frame protocol. Among other parameters the frame protocol needs to cater for HSDPA flow control information, priority queue handling and UE capability information.
1-14
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-15
Version 1 Rev 0
New Channels with HSDPA

The support of HSDPA is based on several new physical channels and one new transport channel.
Transport Channel:
HS-DSCH (High Speed Downlink Shared Channel) The HS-DSCH is the actual transport resource carrying the packet data of the user applications. As it also follows the shortened TTI of 2 ms, it allows for short round trip delay in the operation between NodeB and UE. The 2 ms TTI is short when compared to 10, 20, 40 or 80 ms TTIs supported by Rel. 99 and Rel. 4 transport channels. HS-DSCH describes the physical layer processing by MAC-hs of a HSDPA transport block. Dynamic part: TB size = TBS size {1 to 200 000 bits with 8 bit granularity}; modulation scheme {QPSK, 16-QAM}; redundancy / constellation version {1 8}. Static part: TTI {2 ms for FDD}; type of channel coding {turbo coding}; mother code rate {1/3}, CRC size {24 bits} No semi-static attributes are defined for HS-DSCH.
Physical Channels
High Speed Shared Control Channel (HS-SCCH) The HS-SCCH has a fixed spreading factor of value 128 and is configured only in the downlink direction. It also adopts the shortened TTI of 2 ms. In theory, up to 127 HS-SCCHs can be configured in a cell. However, the UE is required only to be able to listen to up to four HS-SCCH in parallel. The HS-SCCH allows the efficient sharing of one or more HS-PDSCHs among different users. Nevertheless every UE needs to be informed on the DCCH via RRC messages about the specific HS-SCCH-set that it shall monitor in order to receive data via the HS-PDSCHs. High Speed Physical Downlink Shared Channel (HS-PDSCH) The HS-PDSCH has a fixed spreading factor of value 16. Thus, it provides for multicode operation using up to 15 channelization codes in parallel. Of course the UE must support the use of up to 15 channelization codes which depends on its category. The HS-PDSCH adopts the shortened TTI of 2 ms. Uplink Dedicated Control Channel Associated with HS-DSCH Transmission (Uplink HS-DPCCH). The HS-DPCCH has a fixed spreading factor of value 256 and is only configured in uplink direction. The HS-DPCCH also follows the shortened TTI of 2 ms. Its purpose is to provide feedback information about the downlink receive quality and whether the packet data received by the UE are error-free or need to be retransmitted. Thus the NodeB is quickly notified of unsuccessful transmissions and/or changing radio conditions in downlink direction.
1-16
MOTOROLA LTD.2002
Version 1 Rev 0
Transport Channels
Physical Channels
MOTOROLA LTD.2002
1-17
Version 1 Rev 0
HSDPA Basic Operation

The figure consists of three functional parts. The upper part shows the basic communication between UE and NodeB via MAC-hs for packet transmission and retransmission. The lower left part depicts the Uu-interface together with the newly introduced physical channels and transport channel. Note for simplicity reasons the red colored arrows represent the individual HS-DPCCHs per UE. In the lower right corner we sketched a basic scheduling principle illustrating how the HS-DSCH resources can be assigned among several UEs.
MAC-hs
MAC-hs located in NodeB receives user packets from the SRNC. MAC-hs is responsible for transmission and in case of erroneous reception also for retransmission of user packets. The retransmission of user packets is commanded by NodeBs MAC-hs which represents the fundamental change in HSDPA compared to Rel. 99 or Rel. 4. In legacy UMTS releases retransmission are always performed between the RLC peers in UE and SRNC. Of course in downlink the SRNC retransmits the user packets if the UE indicates a Nack. The UE sends feedback information about the downlink channel quality in regular intervals to the NodeB.
New Channels
Two UEs receiving user packets via the HS-DSCH. The physical channels in uplink and downlink are only necessary to firstly signal which UE shall decodea specificHS-DSCH. This is indicated by the UE-id in the HS-SCCH. Secondly after decoding of the user packet, the respective UE has to signal the successful or unsuccessful transmission to the MAC-hs entity in NodeB via the so called Ack/Nack description. HSDPA employs the uplink HS-DPCCH to signal the downlink reception quality and the Ack/Nack description. The downlink reception quality informs the NodeB about the current radio condition which serves as a vital input for NodeBs MAC-hs. The MAC-hs is therefore able to derive the proper modulation scheme and code rate for transmission and retransmission. This process is denoted as AMC in HSDPA.
UE Scheduling
The NodeB also contains a scheduling/priority handling function which determines whether a new transmission or retransmission shall be performed. The green and blue line represent the changing downlink channel quality reported by UE in the uplink. One scheduling method which could be implemented in NodeB is to serve each UE according to the reported downlink channel quality and therefore always exploit best radio conditions. This method maximizes user throughput as it allows to use 16-QAM and an aggressive code rate if C/I is high. Another option is to serve each user proportionally fair despite unfavorable downlink radio conditions. This benefits especially UE 2 which indicates for several TTIs a bad radio quality, but this second method assures at least a minimum guaranteed throughput for UE 2.
1-18
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-19
Version 1 Rev 0
HSDPA and DPCH Operation HSDPA Setup

HSDPA Setup via DPCH
The resource allocation of HSDPA requires the previous setup of a dedicated logical control channel DCCH which is mapped on a dedicated transport channel (DCH). The DCH transport channel runs on a dedicated physical channel (DPCH). Upon RRC connection request sent by UE, the SRNC may request the UEs HSDPA capabilities. The HSDPA related configuration supported in a cell is signaled to the UE via a DCCH mapped onto a DCH which is carried by DPDCH+DPCCH. By decoding the so called "High Speed Information" on the DCCH the UE obtains the knowledge of the physical layer configuration for HSDPA in the Cell. This physical layer configuration allows the UE to decode the HS-SCCH which informs the UE about available user data on the shared transport channel HS-DSCH. Note, that there are no parameters broadcast via BCCH about a cells HSDPA capability. The basic HSDPA setup is explained below: 1. 2. Before moving in CELL_DCH state the UE is told by the RNC to reveal its HSDPA capabilities and category e.g. via RRC Connection Setup message. Once the RNC has obtained the UEs HSDPA capabilities, the subsequent RRC configuration messages (e.g. Radio Bearer Setup) contain the high speed information telling the UE about the HSDPA configuration in the cell. High speed information contains e.g. the UE identity (called H-RNTI) which is implicitly encoded on HS-SCCH to identify the very UE getting HS-DSCH resources allocated, the scrambling code to be applied for HS-PDSCH and HS-SCCH and the HS-SCCH channelization code-set. This is necessary to inform the UE about the decoding of the HS-SCCH which contains specific information about the HS-PDSCHs format carrying the HS-DSCH. From this moment on the UE is in a so called HSDPA standby mode ready to decode HS-SCCHs. Note: The UE might be told to monitor up to four HS-SCCHs. A DPCH must always exist before a UE is able to operate in HSDPA as it carries the DCCH. The DPCH handles the real-time services, e.g. AMR 12.2 voice channel and the RRC signaling via SRBs. On SRBs the necessary RRC signaling messages are exchanged e.g. to reconfigure the physical link for HSDPA or prepare for HS-DSCH cell change. Please remember that HSDPA is not (yet) intended for services with real-time QoS requirements. If there is packet data to be transferred to the UE in the downlink, the NodeB will relay the necessary information for decoding the HS-DSCH to the UE on HS-SCCHs. The UE must be capable of decoding up to four HS-SCCH in parallel. This is a so called HS-SCCH-set. From the HS-SCCHs the UE obtains information how to decode the subsequent HS-PDSCHs and finally obtains the user data from the HS-DSCH. If the UE detects consistent control information intended for it, the UE shall start receiving the HS-PDSCHs. Consistent control information means e.g. that the UE decodes its H-RNTI in one of the assigned HS-SCCHs of the HS-SCCH code-set.
3.
4.
5.
1-20
MOTOROLA LTD.2002
Version 1 Rev 0
HSDPA and DPCH Operation HSDPA Setup HSDPA Setup via the DPCH
MOTOROLA LTD.2002
1-21
Version 1 Rev 0

HSDPA Transmission and Retransmission
6. Whenever there is packet data to be transmitted to a UE, the HS-SCCH indicates to the respective UE that the successive HS-DSCH transport channel contains user data destined for it. If the decoding of the HS-DSCH transport block fails because of CRC error, it is going to report a NACK in the corresponding uplink HS-DPCCH channel. 7. The UE shall transmit acknowledgement information received from the MAC-hs-entity to inform the NodeB whether the HS-DSCH subframe was decoded correctly or incorrectly. 8. If the HS-DSCH was received incorrectly and this result is reported by the UE on uplink HS-DPCCH, the NodeB schedules retransmission in case of RLC-AM. The MAC-hs in NodeB informs the UE with the next possible HS-SCCH whether a retransmission or a new transmission of HS-DSCH transport block will be performed. While the UE is in HSDPA standby or receiving data, it may send CQI reports about the downlink reception quality to the serving NodeB. Physical Channel Combinations supported with HSDPA on the same FDD Frequency Physical Channel Combination Transport Channel Combination Mandatory or dependent on UE radio access capabilities Depending on UE radio access capabilities Comment
Uplink
DPCCH + one or more DPDCHs + HS-DPCCH
One or more DCHs coded into a single CCTrCH
The maximum number of DCHs and the maximum bit rate are dependent on UE radio access capabilities. This combination is required in case HS-DSCH(s) are configured. The maximum number of DCHs and the maximum channel bit rate are dependent on UE radio access capabilities This combination is required by a UE in CELL_DCH state to be able to read the SFN of a neighbouring cell and support "SFN-CFN observed time difference" and "SFN-SFN observed time difference" measurements while HS-DSCH(s) are configured.
Downlink
DPCCH + one or more DPDCH + one or more HS-SCCHs + zero, one or more HS-PDSCHs PCCPCH (neighbour cell) + DPCCH + one or more DPDCH + one or more HS-SCCHs + zero, one or more HS-PDSCHs
One HS-DSCH coded into a single CCTrCH + one or more DCH coded into a single CCTrCH BCH (neighbour cell) + one or more DCHs + one HS-DSCH
Depending on UE radio access capabilities
Depending on UE radio access capabilities
1-22
MOTOROLA LTD.2002
Version 1 Rev 0
HSDPA and DPCH Operation HSDPA Setup HSDPA Transmission and Retransmission
MOTOROLA LTD.2002
1-23
Version 1 Rev 0
HSDPA High Speed Information

When the UE is in CELL_DCH state, RRC-messages are exchanged e.g. for setting up radio bearers. If the UE has indicated beforehand its HSDPA capability, the RB_SETUP-message may contain HSDPA related information such as:
H-RNTI
The H-RNTI (High Speed Radio Network Transaction Identifier) is firstly used to identify the UE which shall receive data from a HS-DSCH. Therefore the UEs H-RNTI is implicitly encoded in the HS-SCCH. Implicitly encoded means that part 1 of HS-SCCH uses H-RNTI as input for UE specific masking. For part 2 of HS-SCCH, the H-RNTI is used for UE specific CRC attachment. Thus the UE is able to distinguish which out of a maximum of four HS-SCCHs contains decoding information for the following HS-PDSCHs in the sub-sequent HS-DSCH subframe. This means on the other hand, that the actual HS-DSCH TB does not contain any UE-id. Thus the HS-DSCH resource allocation in HSDPA is entirely done without higher layer involvement. The reasons for this are twofold. Leaving the resource assignment at layer 1 speeds up the entire process (avoiding higher layer processing delays) and avoids to transmit UE specific RNTIs and sequence numbers. The timely relationship between HS-SCCH and HS-DSCH avoids the need of attaching an UE-id within HS-DSCH transport block to identify the recipient. For each HS-DSCH TTI of 2 ms, each related HS-SCCH occurring 2 slots ahead of HS-DSCH carries HS-DSCH related downlink signaling for one UE only. The H-RNTI encoded in the HS-SCCH identifies therefore an UE having a HS-PDSCH assignment within a cell. Therefore the UE needs at first to decode the HS-SCCH (part 1) before it can attempt to demodulate the allocated HS-PDSCHs. The HS-SCCH signaling message is divided into two parts, with part 1 containing the time critical information on channelization code set and modulation scheme. Part 2 of HS-SCCH consists of transport block size and HARQ-related (indicating new transmission or retransmission, HARQ process id, redundancy and constellation version). A 16-bit UE specific CRC is computed over part 1 and part 2 and attached to part 2. The UE ID is only implicitly included in the CRC. In order to allow the UE quickly to determine whether HS-PDSCHs are going to be allocated, the part 1 alone of HS-SCCH is scrambled with the H-RNTI as well. If more than one UE shall be served within a subframe, more than one HS-SCCH needs to be transmitted as one HS-SCCH can only allocate HS-DSCH resources for one UE only. However code multiplexing may be limited between 2-3 users per subframe as it becomes less efficient and processing is quite high for NodeB.
1-24
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-25
Version 1 Rev 0
DL-HSPDSCH-Info
DL-HSPDSCH-Info
The downlink HS-PDSCH information consists of HS-SCCH Info and measurement feedback info. (Measurement feedback info will explain at a later stage in chapter 3.)
HS-SCCH Info
Downlink Scrambling Coding This information element informs the UE about the DL scrambling code to be applied for HS-PDSCHs and HS-SCCHs. Default is same scrambling code as for the P-CPICH. If dl-Scrambling Code IE is present, then secondary scrambling code is used in the cell under which HSDPA operates. HS-SCCH Channelization Code info This information element contains the maximum number of HS-SCCH codes that can be assigned to a UE. In Rel. 5 this parameter is set to four. HS-SCCH Codes This information element tells the UE the specific channelization codes used for the HS-SCCH set. As a HS-SCCH channel has a fixed spreading factor of 128, the channelization code ranges from 0127. The channelization code C(ch,128,0) cannot be used as the related branch is already occupied by the P-CPICH and P-CCPCH. The same is valid for C(ch,16,0). And also if secondary scrambling codes are used. NOTE HSDPA may operate on a different frequency than legacy UMTS traffic so almost the entire resources of the second carrier could be dedicated to HSDPA.
1-26
MOTOROLA LTD.2002
DL-HSPDSCH-Info
Version 1 Rev 0
DL-HSPDSCH-Info
MOTOROLA LTD.2002
1-27
Version 1 Rev 0
Channel Type Switching with HSDPA

The figure shows only the important RRC connected states of a UE in UTRAN. We would like to show here that even with HSDPA transport channel type switching is feasible which saves resources in UTRAN and battery in UE. A prerequisite is that HSDPA retains MAC-d in SRNC. In order to ease the understanding we distinguish the Rel. 99 CELL_DCH state between CELL_DCH & HSDPA and CELL_DCH ONLY. Note the CTS description does not intend to be exhaustive. For detailed information about the various RRC states and transitions between them the reader kindly is referred to [8].
CELL_DCH ONLY
This RRC state is identical to the CELL_DCH in Rel. 99. Here we would like to highlight that whenever the UE moves from a HSDPA capable cell to a non-capable cell, the radio bearer(s) have to be reconfigured to e.g. a ordinary 384 kbit/s bearer known since Rel. 99.
CELL_DCH & HSDPA

Here the UE is actually in HSDPA standby and ready to receive data on HS-DSCH. A DCCH and possibly also DTCHs (e.g. for real time traffic) exist and are transmitted via DCHs. Soft and softer handover are possible with DCH. UE is known on Cell level UE has been configured for HSDPA (e.g. through Radio Bearer Setup or Physical Channel Reconfiguration message) and is monitoring the assigned HS-SCCH code-set. For HSDPA there are DTCHs configured and ready to use. High speed user data is transmitted to the UE within 2 ms subframes. If the UE changes from a HSDPA capable cell to a cell without HSDPA capability, then the packet switched radio bearers need to be reconfigured to e.g. a 384 kbit/s bearer. This allows user data exchange within the packet domain to continue. Changes of the HSDPA serving cell (serving HS-DSCH radio link change) are performed through hard handover through e.g. Transport Channel Reconfiguration or Radio Bearer Reconfiguration messages. Of course, if the UE changes back to a HSDPA capable cell, bearer reconfiguration puts the UE back into HSDPA standby. If the UE is idle (not receiving any user data on HS-DSCH) for a while being in Cell_DCH with HSDPA setup, the SRNC may put the UE into CELL_FACH state.
CELL_FACH
The advantage of being in CELL_FACH state is twofold. Firstly the UE does not have to monitor the HS-SCCH code-set anymore which extends its battery capacity. Secondly the UTRAN saves channelization code resources as the UE does not longer occupy DCHs. Through traffic volume measurements the UE reports its RLC transmit buffer occupancy and a rise in uplink or downlink (RLC transmit buffer in SRNC) above a certain threshold triggers the reconfiguration of UE into CELL_DCH and possibly into HSDPA standby mode.
1-28
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-29
Version 1 Rev 0
HSDPA Control and User Plane Extensions

Control Plane Extensions:
New NBAP, RRC & RNSAP IEs Theses protocols are extended in Rel. 5 with new IEs in order to support the HSDPA procedures like e.g. Intra-NodeB Serving HS-DSCH Radio Link Change, HS-DSCH MAC-d Flow Addition/Deletion etc. Examples are: the AUDIT procedure is extended by the HS-DSCH resources Information IE for every cell which has been configured with HS-DSCH resources. RRC messages contain new information elements like for example the HS-DPCCH feedback cycle IE in uplink. RNSAP-messaging must also be extended similar to NBAP to cope with the additional HSDPA functionality. Radio Resource Sharing Radio Resource Sharing has to be extended in order to support the shared operation of dedicated and shared channels in a cell. For example, a simple algorithm can be used for best effort services in HSDPA to always use the capacity (e.g. available transmission power) that remains unused by the DCH channels in a cell for HS-DSCH and HS-SCCH. However, streaming services require an algorithm with a dynamic behavior when sharing dedicated and HSDPA channels to assure a constant throughput. Both transport channel types HS-DSCH and DCH share the available downlink transmission power in the NodeB. Admission / Congestion Control The additional transmission output power and higher throughput rates have to be taken into account by the admission control algorithm in order to protect existing services from quality degradation due to too high interference generated by HSDPA transmission. Updated thresholds for overload defense strategies have to be implemented. Ultimately, congestion control has to be updated with new thresholds and mechanisms for denying services, downlink power up commands, throughput reduction and last but not least deliberate dropping if the overall QoS is jeopardized. A vendor dependent solution can be assumed for such capacity control issues in conjunction with HSDPA support.
User Plane Extensions:

PDCP/RLC/MAC-d These user plane protocols have to be (re-)designed to meet the higher bandwidth needs for user allocated to HSDPA channels. PDCP may perform header TCP/IP header compression or not, RLC parameters (mode, timers, TTI etc.) are configured like in Rel. 99 / Rel. 4 but now much higher throughput rates must be supported by processing boards in the RNC. Frame Protocol A new variant of frame protocol, called HS-DSCH FP, has to be implemented with a suitable flow control in order to support the higher throughput rates. The TTI of RLC is of Rel.99 / Rel.4 (10 ms, 20 ms, 40 ms or 80 ms) whereas the TTI of MAC-hs is 2 ms.
1-30
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-31
Version 1 Rev 0
HSDPA Packet Switched Protocol Stacks

The protocol stack of HSDPA can be divided like it is done for legacy UMTS releases in control plane and user plane function. Note that HSDPA is only meant to operate in the packet domain of UMTS because of the use of shared resources preventing real time traffic. However, real time service could be handled in parallel via the circuit switched domain beside HSDPA. Note that the transport layer in UTRAN might be based solely on IP only in Rel. 5.
HSDPA Control Plane

The packet-switched control plane is used for the exchange of control information related to packet-switched services between UE and UTRAN as well between UE and Core Network. In addition, the packet-switched control plane can be used for the exchange of short messages. With the introduction of HSDPA the Rel. 99 / Rel. 4 protocol stack remains nearly unchanged except that a new MAC-hs (high speed MAC) is introduced in NodeB. Access Stratum Protocols The Access Stratum (AS) protocols are all protocols within the rectangle that is defined by the RRC-protocol (Radio Resource Control) at the upper left corner, the MAC-protocol (Medium Access Control) at the lower left corner and the RANAP-protocol (Radio Access Network Application Part) at the upper right corner. Non Access Stratum Protocols The Non Access Stratum (NAS) protocols are GMM (GPRS Mobility Management), SM (Session Management), and GSMS (SMS over packet-switched domain). Note that for HSDPA, the control plane signaling has to be extended in order interact and inter-connect with MAC-hs. HSDPA brings new IEs and procedures (e.g. HS-DSCH cell change) thus NBAP, RNSAP and RRC signaling have to be extended. MAC-hs The high speed MAC controls the operation of the shared resources through new the signaling control channels HS-SCCH and HS-DPCCH. Please note that the HS-SCCH is a point to multipoint control channel received by all UEs which are in HSDPA standby and having the same HS-SCCH channelization codes assigned. On the other side the HS-DPCCH is a point to point signaling channel which is unique per UE as it is scrambled by UEs uplink scrambling code. MAC-hs is located in NodeB and gets configured for each UE through NBAP messages only extended by HSDPA specific IEs (Rel. 5 information elements). The SRNC setups HSDPA in UE via RRC signaling messages. Both UE and SRNC communicate via Rel. 99 / Rel. 4 RRC messages which are extended by HSDPA specific IEs (Rel. 5 information elements).
1-32
MOTOROLA LTD.2002
Version 1 Rev 0
HSDPA Packet Switched Protocol Stacks HSDPA Control Plane Protocol Stack
MOTOROLA LTD.2002
1-33
Version 1 Rev 0

HSDPA User Plane
Configuration without MAC-c/sh In this case, the CRNC does not have any user plane function for the HS-DSCH. MAC-d in the SRNC is located directly above MAC-hs in the NodeB, i.e. in the HS-DSCH user plane the SRNC is directly connected to the NodeB via the HS-DSCH Frame Protocol, thus bypassing the CRNC. MAC-d entity UE Side MAC-d has a connection to the MAC-hs entity. This connection is used to receive data from the HS-DSCH transport channel which is handled by MAC-hs. MAC-hs entity UE Side MAC-hs handles the HSDPA specific functions like HARQ, data queue distribution, reordering of MAC-hs PDUs and disassembly of MAC-hs PDUs. MAC-d entity UTRAN Side MAC-d covers the following functionality: transport channel type switching, channel type field multiplexing, ciphering/deciphering for RLC TM mode, DL scheduling / priority handling and flow control. A flow control function exists towards MAC-hs transparently passing through a CRNC, if present. MAC-d is responsible for the mapping of dedicated logical channels onto the available dedicated transport channels or routing the data received on a DCCH or DTCH towards MAC-hs. MAC-hs UTRAN Side There is one MAC-hs entity in the NodeB for each cell that supports HS-DSCH transmission. The MAC-hs is responsible for handling the data transmitted on the HS-DSCH as well for the management of the physical resources (HS-PDSCH, HS-SCCH, HS-DPCCH) allocated to HSDPA. There is a companion flow control function between MAC-hs and MAC-d. Both entities provide a controlled data flow taking the transmission capabilities of the air interface into account in dynamic manner. This reduces discarded and retransmitted data as a result of HS-DSCH congestion or HS-DSCH cell change in RLC-UM (less RL-UM PDUs have to be discarded in case of HS-DSCH cell change). Note: If CRNC and SRNC are identical, then actually nothing changes for MAC-d and NodeBs MAC-hs except that the Iur interface is not present.
1-34
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-35
Version 1 Rev 0
HSDPA Timing Relations

HSDPA Downlink Timing
The HSDPA HS-SCCH is aligned with the P-CPICH and the P-CCPCH on which the cell SFN (System Frame Number) is transmitted. The HS-DSCH follows the HS-SCCH with a fixed offset. The HS-PDSCHs carrying the HS-DSCH is/are always transmitted with a fixed offset of 5120 chips after the related HS-SCCH subframe has started. Note that the users Rel. 99/Rel. 4 DPCH and HS-SCCH/HS-DSCH are not slot aligned but differ always by a multiple of 256 chips in order to maintain downlink orthogonality. Therefore some additional timing constraints have to be considered in uplink. DPCH (Dedicated Physical Channel) The DL-DPCH consists of DPDCH and DPCCH which are time multiplexed. The DPDCH is the physical bearer for the DCH transport channel which carries DCCH and may be also DTCH (e.g. real time services). The DPCH timing may be different for different DPCHs, but the offset from the CPICH frame timing is always a multiple of 256 chips, i.e. t(DPCH,n) = Tn x 256 chips, Tn = {0, 1, 2, 3 149} resulting in maximum chip offset of 38,144 chips. The DPCH is divided into 15 slots per radio frame allowing for a fast power control in every slot (1500 times / s) via DPCCH. HS-SCCH (High Speed Shared Control Channel) The HS-SCCH is time aligned with the CPICH. It has a fixed spreading factor of 128 which results in a fixed rate of 2 x 38400/128 = 60 kbit/s. The HS-SCCH is transmitted in downlink only and is used to carry HS-DSCH related signaling information. In difference to the DPCH, the HS-SCCH is divided into five subframes per radio frame resulting into a duration of 2 ms per subframe. One subframe consists therefore of three slots resulting in 2560 chips/slot x 3 = 7680 chips HS-PDSCH (High Speed Physical Downlink Shared Transport Channel) The HS-PDSCH is the physical bearers for the HS-DSCH. The HS-PDSCH has a fixed spreading factor of 16. Like the HS-SCCH the HS-DSCH is transmitted in subframes. However the start of HS-DSCH subframe is always delayed by 5120 chips compared to the related HS-SCCH allowing the UE to obtain the most important information on HS-SCCH before attempting to decode the very HS-DSCH subframe. Note: The HS-SCCH subframe structure is aligned with the CPICH. Default is the P-CPICH. The downlink DPCH timing is not tied to the HS-SCCH or consequently HS-DSCH timing.
1-36
MOTOROLA LTD.2002
Version 1 Rev 0
HSDPA Timing Relations HSDPA Downlink Timing
P-CPICH
HS-PDSCH
HS-PDSCH
HS-PDSCH
HS-PDSCH
HS-PDSCH
MOTOROLA LTD.2002
1-37
Version 1 Rev 0

HSDPA Uplink Timing
It can be seen that the HS-DPCCH is not slot aligned with uplink DPCH of Rel. 99/Rel. 4. However, the correct timing to start transmission on HS-DPCCH must be derived from the timing of the uplink DPCH, to ensure uplink orthogonality with uplink DPCCH. For simplicity reasons we provide here only a simplified drawing. Further details are given in chapter 2. The HSDPA operation procedure has strictly specified timing values for the UE operation starting from HS-SCCH reception via HS-DSCH decoding to the uplink Ack/Nack transmission on HS-DPCCH. 7.5 slots delay of uplink HS-DPCCH The key timing from UE point of view is the approximately 7.5 slots (= 7.5 x 2/3 ms = 5 ms) from the end of the HS-DSCH TTI to the start of the Ack/Nack transmission on the HS-DPCCH in the uplink as can be seen in the figure. To aceive this minimum 5ms and orthogonallity with the other uplink channels, the HS-DPCCH is time synchronised to start m x 256 chips after the start of the UL DPCCH, which in turn starts T0 chips after receipt of the DL DPCH. m is derived from the time difference between the downlink Rel 99/Rel 4 DPCH and the start of the respective HS-PDSCH sub-frame Therefore the uplink HS-DPCCH may start in the middle of the uplink slot and this needs to be taken into account in the uplink power setting process ensuring a proper decoding in the uplink in NodeB. The UE capabilities do not impact the timing of individual TTI transmission but do define how often the NodeB can schedule transmission to a particular UE. The UE capabilities include information about the minimum inter-TTI interval that tells whether consecutive TTIs may be used or not for transmitting HS-DSCHs to the same UE. Thus this provides a relieve for certain UEs which do not have to receive in downlink in consecutive subframes neither transmit in the uplink on consecutive subframes. NOTE TTI interval in HSDPA always refers to the 2 ms subframe duration
1-38
MOTOROLA LTD.2002
Version 1 Rev 0
HSDPA Timing Relations HSDPA Uplink Timing
n-th DPCH: DPDCH/D + DPCCH/D (SF = 4.....512) HS-PDSCH m HS-PDSCH HS-PDSCH HS-PDSCH HS-PDSCH
m x 256 Chips
HS-PDSCH
T0
MOTOROLA LTD.2002
1-39
Version 1 Rev 0
Composition of the HS-SCCH

As illustrated in the figure, the HS-SCCH is split into two parts, part 1 and part 2 which are independently coded from each other. Part 1 carries the TFRI (transport format and resource related information. Part 2 carries Hybrid ARQ related information. Every HS-SCCH subframe employs UE specific masking with the H-RNTI for both parts allowing the UE to determine whether the detected control channel is actually intended for this particular UE. It can be seen that the part2 of HS-SCCH carries a CRC which is build over part1 and part2 information together.
Modulation Type and Number of HS-PDSCHs

HS-PDSCH channelization-code informs UE about how many HS-PDSCHs in parallel shall be decoded. Modulation scheme informs UE if QPSK or 16-QAM is employed for HS-PDSCH.
HARQ Info (Hybrid-ARQ-related Information)

Transport block size used for HS-DSCH. HARQ protocol related information for the corresponding HS-DSCH TTI and information about the redundancy version necessary for retransmission combining. HARQ parameters inform the UE if the HS-DSCH subframe contains either a new TB or a repetition of a Nacked TB and how original transmission and retransmission have to be combined based on redundancy parameter. Note: The HS-PDSCH subframe can be de-spread and demodulated by the UE without having to wait for complete reception of the preceding HS-SCCH subframe.
1-40
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-41
Version 1 Rev 0
Hybrid ARQ Techniques

ARQ (automatic repeat request) is basically a mechanism which allows the receiver to request retransmission of erroneous received packets. Hybrid ARQ schemes (HARQ) can be classified as type I, type II and type III depending on the level of complexity employed in their implementation.
Type I Hybrid ARQ Low UE Complexity

The ARQ method used in Rel. 99 / Rel. 4 specifications is referred to as HARQ type I. In this basic HARQ type I, a CRC is added to the TB and the data is encoded with a forward error correction (FEC) code. In the receiver the FEC code is decoded and the quality of the packet is checked (CRC check). If there are errors in the packet, a retransmission of the packet (RLC-PDU) is requested. The erroneous packet is discarded and retransmission uses the same coding as the first transmission. In EGPRS for instance LA (link adaptation) is used beside IR for retransmission of erroneous packets. LA means adjusting the Modulation and Coding Scheme (MCS) for retransmission. Basically a more robust MCS could be used to increase the likelihood of successful transmission.
Type II Hybrid ARQ High UE Complexity

In this HARQ scheme, also denoted as full IR (incremental redundancy), retransmission requests only consist of parity bits. The original transmission must be self-decodable, but the retransmission are not and typically not identical with the original transmission. The receiver combines additional parity bits from retransmission with bits of the first transmission resulting in lower code rates, before decoding is attempted. The basic difference between type I and type II is that when retransmitting a coded block with HARQ type II, each retransmission is punctured using a different puncturing pattern than the original coded block. Thus the level of redundancy is increased for each retransmission. Type III is a variant of type II wherein each retransmission is self-decodable.
Type III Hybrid ARQ medium UE Complexity

Like type II, type III hybrid ARQ also belongs to the IR ARQ schemes. This means that retransmission concerning one PDU are not discarded but kept at the receiver for combination with additional information before decoding. In situations where the original transmitted PDU is severely damaged, for example, due to interference, it is desirable to have a scheme where any additional information sent is self decodable. In type III HARQ each retransmission is self-decodable but different redundancy bits are retransmitted in various transmissions of the same PDU. Thus, the data can be recovered from the retransmitted packet even without combining with previous transmission instances if it is transmitted with sufficient quality. Chase combining is yet another special form of type III HARQ where each transmission instance contains the same bits like in the original transmission.. Note: Both methods chase combining and IR perform combining of original transmitted and retransmitted packets based on soft decision bits. This means the combining is made by simply adding soft bits after equalization. Chase combining involves the retransmission of the same coded block in each retransmission, whereas in IR, instead of simple repetitions, additional redundancy (parity bits) information is transmitted with each retransmission until the base code rate is reached.
1-42
MOTOROLA LTD.2002
Version 1 Rev 0
Type I
Type II
Type III
L1 Soft Combining
Link Adaptation
Incremental Redundancy
Full IR
Partial IR
Chase Combining
Possible Retransmission Methods
MOTOROLA LTD.2002
1-43
Version 1 Rev 0
DCH / HS-DSCH Comparison

Variable Spreading Factor
The SF does not vary for the DCH on a frame by frame basis but is chosen upon setup according to the max. expected data rate. The data rate variation on the DCH during operation is taken care of through either rate matching or through DTX. Compared to that, the SF for the HS-DSCH is always fixed at SF=16, the data rate variation is achieved by AMC and therefore no explicit RRC signaling is needed.
Fast Power Control

The DCH DL power control is based on the inner-loop power control commands sent by the UE on the associated uplink DPCCH. Since the HS-DSCH uses AMC, fast power control in downlink is on the one hand not allowed when 16-QAM is used. Note, 16-QAM requires amplitude estimations by the receiver. On the other hand it is not needed as the DL power control dynamics is quite limited (about 20 dB) due to intra-cell interference and by NodeB implementation. Intra-cell interference occurs between users on parallel code channels. This interference is caused by various propagation delays in downlink. Thus former transmitted orthogonal channelization codes at NodeB side are no longer orthogonal when they arrive at the UE. For a user close to NodeB, the power control cannot reduce power maximally, and on the other hand reducing the power beyond 20 dB dynamics would have only marginal impact on the DCHs capacity. With HSDPA, this property is now utilized by the link adaptation function and AMC to select a coding and modulation combination that requires higher Ec/Ior, which is only available for user close to the cell. To enable a large dynamic range of the HSDPA link adaptation and to maintain a good spectral efficiency, a user may simultaneously utilize up to 15 HS-DSCH channelization codes in parallel. The use of more robust coding, Hybrid ARQ and multicode operation removes therefore the need for a variable SF and fast power control.
AMC
The MAC-hs in NodeB is applying adaptive modulation and coding only for the HS-DSCH. Note, the detailed function of AMC is explained in chapter 2.
Multicode Operation
Both HS-DSCH and DCH may use more than one channelization code, but for HS-DSCH the UE can go for the maximum of either 5, 10 or 15 parallel channelization codes with SF = 16 depending on the UEs category. Therefore HSDPA uses the code multiplexing extensively compared to DCH operation.
Fast Layer 1 ARQ

The significant difference between DCH and HS-DSCH is that MAC-hs allows for retransmission in NodeB directly at the air interface once the UE indicates a negative reception via the uplink DPCCH feedback channel. Therefore the higher layer RLC located in SRNC is not confronted with retransmission in the first place, whereas for DCH operation, RLC takes care of retransmission. Note: retransmission isonly performed in RLC acknowledged mode.
1-44
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-45
Version 1 Rev 0
DPCH / HS-PDSCH Comparison contd

16-QAM Modulation / Interleaving / Channel Coding Scheme
DPCH uses the Rel. 99 QPSK modulation with interleaving depth between 10-80 ms depending on the configured TTI. Turbo and convolutional coding can be configured for the DCH, whereas HS-PDSCH exclusively uses turbo coding. In order to reduce the latency for retransmission, HS-PDSCH uses a shortened TTI of 2 ms and can interchange between QPSK and 16-QAM modulation depending on UE capability and radio conditions.
TrCH Multiplexing
The HS-DSCH is a transport channel with only one TB (transport block) per TTI per UE, so there is no need for static rate matching, transport block concatenation or balancing of quality. This complex matter shall be explained here in detail now:
Static rate matching - DCH versus HS-DSCH

On transport channels, transport blocks are exchanged between MAC layer and the physical layer. For each transport block there is a CRC checksum calculated which provides for error detection. After that, channel coding is performed. After channel coding the code blocks become encoded blocks and are treated in rate matching. Static rate matching ensures, that there are firstly only as many coded bits to be transmitted as there are bits available on the physical channel but also it balances the coding rate between different transport channels according to their relative importance. In rate matching, the bits of encoded blocks are repeated or punctured according to the available bandwidth and according to the RMAs (rate matching attributes). Those RMAs define the priorities between the coded bits of different transport channels, and are set by layer 3 (semi-static attribute). For instance, by setting RMA of transport channel x to twice the value of the RMA of a transport channel y, the coded bits of the TrCH x are made twice more important than the coded bits of TrCH y. This mechanism allows for balancing the coding rate between transport channels. As HSDPA uses only one transport channel per TTI, there is no need to match this only one transport channel to the capacity of the available physical channels. HSDPA uses a fixed SF = 16 for HS-DSCH and there is only one transport block per TTI, thus the entire transport block can be mapped onto the HS-DSCH. Only one transport channel per TTI is allowed in HSDPA, so there is never the case that two transport channels of type HS-DSCH must be mapped onto a CCTrCH and need to apply rate matching.
Transport block concatenation - DCH versus HS-DSCH

In HSDPA there occurs only 1 transport block per HS-DSCH TTI, thus no case where 2 or more transport blocks need to be mapped onto the same HS-DSCH. As HSDPA uses only transport channel, the balancing of quality is not applicable as quality balancing makes only sense if more than one TrCH are in use.
Soft(er) Handover
For HS-DSCH there is no macro diversity possible as it is a common transport channel. Of course DCH supports SHO.
1-46
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-47
Version 1 Rev 0
Advantages and Disadvantages of HSDPA

Data services are anticipated to have an enormous growth rate over the coming years and will likely become the dominating source of traffic load in 3G mobile cellular networks. As more sophisticated services evolve, a major challenge of cellular systems design is to achieve a high system capacity and simultaneously facilitate a mixture of diverse services with very different QoS-requirements. HSDPA tries to meet those requirements and brings quite significant benefits for the subscribers:
Advantages of HSDPA
Reduced Latency for retransmission With MAC-hs installed in NodeB, the uplink feedback channel and the shortened TTI of 2 ms, reception of erroneous packet data results in a faster retransmission. On top of this, no transmission is wasted due to the use of incremental redundancy or chase combining. Higher DL Throughput and Peak Throughput Rates up to 14 Mbit/s As HSDPA makes extensive use of code multiplexing in conjunction with higher order modulation, the downlink throughput can be increased beyond 10 Mbit/s. If the radio conditions are favorable, AMC allows a very aggressive code rate of 1:1, thus offering peak data rates up to 14 Mbit/s. Support of Non-Real Time and Streaming Applications Through a sophisticated packet scheduler implementation in the NodeB it is possible to manage stringent QoS requirements of streaming applications and of course less time critical services as well. The packet scheduler is able to adjust the code rate, modulation type, number of channelization codes and the transmission power within every subframe. Thus the NodeB is highly flexible in assuring various QoS requirements of subscribers services. High Spectrum and Code Efficiency due to 16-QAM Spectrum efficiency increases when the modulation scheme changes from QPSK to 16-QAM as literally more kbit/s respectively Mbit/s can be transmitted in the 5 MHz bandwidth. Code efficiency increases when the UE is able to dynamically make use of the available channelization codes. In downlink, the WCDMA power control dynamics is in the order of about 20 dB. It is limited by the intra-cell and inter-cell interference compared to the uplink power control dynamics of about 70 dB. This means that for a particular user, very close to the NodeB, the DL power control cannot reduce the power maximally, and on the other hand reducing power beyond 20 dB dynamics may only have marginal impact on the downlink capacity. HSDPA now utilizes this available DL capacity though link adaptation and through AMC. The higher Ec/Io signal quality (when user is close to NodeB) is exploited by selecting a very aggressive code rate and 16-QAM. Vice versa, if radio conditions become worse, HSDPA switches down to the more robust QPSK, applies lower code rate (meaning more protection) but may use increasing number of channelization codes to support the users QoS demands (e.g. guaranteed throughput rate).
1-48
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-49
Version 1 Rev 0
Disadvantages of HSDPA
Higher Processing power in UE
The higher throughput rates in the downlink require that the UEs processors and DSPs are designed accordingly. UEs battery consumption will increase, too. This imposes serious challenges on the UE vendors.
More Memory in UE due to HARQ

The HARQ is making use of every previous transmission instance of a data packet. In case of incremental redundancy, the UE is required to buffer previously transmitted data blocks for soft combining. Compared to that, chase combining allows the UE to add the soft decision values after every transmission instance of a data packet, thus less memory is needed. Incremental redundancy requires the maximum UE memory. Therefore the terminal default memory requirements are set on the basis of chase combining at the maximum data rate which is supported by the terminal. Hence, at the highest data rate, only chase combining may be used, while with lower data rates, also incremental redundancy can be used. The UE category value informs the RNC about the so called "total number of soft channel bits" which may limit the IR performance if it comes to peak data rates.
Advanced Receiver for 16-QAM

With QPSK only a phase estimate is necessary for the demodulation process. However, with 16-QAM an amplitude estimation is additionally required to separate the constellation points on the I-Q plane. Furthermore, accurate phase information is needed since constellation points have smaller decision space in the phase domain compared to QPSK.
Higher Vulnerability using 16-QAM

For peak data rates using 16-QAM, higher C/I (Eb/No) values are required and thus the HS-DSCH is more vulnerable to interference including inter-symbol interference. As 16-QAM needs both the phase and the amplitude to determine the symbol value, amplitude distortions have big impact. Compared to that in QPSK where only the phase is considered for determining the symbol value, amplitude distortions are negligible. Beside this, the HSDPA transmitters with 16-QAM need an improved linearity of the power amplifier chain in terms of EVM performance (Error Vector Magnitude).
Software and possibly Hardware Upgrade

Depending on the very UTRAN vendor, a hardware upgrade might be necessary for the NodeB in order to support 16-QAM. Additional software upgrades in NodeB and RNC complement the HSDPA support.
1-50
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-51
Version 1 Rev 0
No Macro-Diversity for HS-PDSCH, HS-SCCH
As HS-PDSCH and HS-SCCH are shared channels, no soft handover is possible and thus the performance decreases particularly at cell edge. Nevertheless mobility procedures like intra-NodeB and inter NodeB cell changes are supported which will mitigate this disadvantage.
Power / Interference Increase in uplink and downlink

HSDPA operation may require quite large transmission power in downlink in order to operate at reasonable BLER (considering 1st transmission) and may therefore increase the inter cell interference. Especially HSDPA operation at cell edge is only feasible at high downlink transmission power for HS-SCCHs and HS-PDSCHs. It should be noted, that the amount of HS-SCCHs increases with code multiplexing of HS-PDSCHs and can add considerably to the downlink interference when users are subjected to unfavorable radio conditions. In the uplink the HS-DPCCH may increase uplink interference if e.g. operated at maximum power value. Proper decoding of HS-DPCCH is on the one hand necessary to avoid unnecessary retransmission but on the other hand high uplink HS-DPCCH transmission power increases the interference among the various users.
Signalling Overhead caused by HS-SCCH (dl) and DPCCH (ul)

The necessary signalling in order to support HS-DSCH consumes valuable channelization code resources. In downlink a spreading factor of 128 is needed per HS-SCCH and in uplink a spreading factor of 256 is needed per UE. Of course, the channelization code consumption in uplink is negligible as every UE has its dedicated uplink scrambling code so there is almost no resource limitation. However, in downlink every HS-SCCH occupies a spreading factor of 128 which is also necessary for one AMR 12.2 kbit/s speech conversation.
1-52
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-53
Version 1 Rev 0
Future Enhancements of HSDPA

UMTS UTRA FDD aims to support a variety of multiple antenna transmission techniques in order to enhance coverage, system throughput and spectral efficiency of HSDPA. A major aim of using multiple antenna transmission in macro-cellular environments is to increase the coverage ratio at medium and higher data rates, lets say 2 Mbit/s and beyond. In a typical deployment, two to four or more transmit antennas might be used per sector.
Beamforming
Beamforming makes use of adaptive antennas and can therefore provide a better C/I to UEs in the downlink. At the same time beamforming allows re-use of scarce downlink channelization codes as the individual UEs are separated in space and possibly through different downlink scrambling codes, thus making use of secondary scrambling codes. The signals toward different UEs from the same cell are typically transmitted under the same primary scrambling code and separated by means of orthogonal channelization codes. However, some of the beams may be transmitted under a secondary scrambling code with its associated channelization code tree, thereby increasing the resources in the cell. Note that the loss of the reduced orthogonality between primary and secondary scrambling code can be partly mitigated in the case of beamforming by splitting the cell into multiple scrambling code regions, so the spatial isolation between beams using different scrambling codes helps to compensate the lack of orthogonality.
Transmit Diversity
The downlink capacity could be improved by using receive antenna diversity in the UE. However for small and cheap mobiles it is not feasible to use two antennas and receiver chains. Therefore, the WCDMA standard already supports the use of base station transmit diversity in Rel. 99. There are two modes: open loop (TSTD and STTD) and closed loop mode (mode1 with phase adjustment only and mode 2 with phase and amplitude adjustment). The open loop mode simply transmits the coded information from two antennas, but on the diversity antenna the bits are time reversed and complex conjugated. The STTD method provides two kinds of diversity. The physical separation of the antennas provides space diversity and the time difference derived from a bit-reversing process provides for time diversity, thus the decoding in the receiver becomes more reliable. The closed loop mode can only be applied to the downlink channel, if there is an associated uplink channel. Thus this mode can only be used with dedicated channels (DPCH, PDSCH or HS-PDSCH with an associated uplink DPCCH).
MIMO
With MIMO (Multiple Input Multiple Output) at the transmitter, x independent data streams are transmitted out of the x antennas on the same frequency band. At the receiver, each antenna receives all of the transmitted sub-streams superimposed, not separately. If multipath scattering is sufficient, these x data streams have different spatial signatures to each of the e.g. p receive antennas and they are separable, the signals arrive with different phases. When a transmitter has x antennas and the receiver has p antennas, the link speed increases linearly with min (x,p) given the same power and bandwidth budget.
1-54
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-55
Version 1 Rev 0
Preview to HSUPA
Preview to HSUPA
The aim of the HSUPA is to enhance the uplink DCH operation and performance, using several techniques in order to support services like video-clips, multimedia, e-mail, telematics, gaming and video-streaming. A study showed that various techniques such as Node-B controlled scheduling, shorter TTI and a hybrid ARQ layer in the Node-B can enhance the uplink packet transfer performance significantly compared to Release-99/Rel-4/Rel-5. The targeted improvements are a higher air interface capacity in the uplink and a better end user experience by increasing both the cell throughput and the coverage of higher bit rates in uplink. There are fundamental differences between uplink and downlink data transmission, making it impossible to simply introduce the same HSDPA solutions for the uplink.
Total available Transmission Power

The key difference between uplink and downlink is the handling of the total transmission power available. In downlink, the power is centralized whereas in the uplink the power available for an individual UE is limited by the terminal amplifier capabilities. Therefore it can be said, that a pure time division approach, being in place with the maximum data rate in HSDPA, would not work for the uplink.
Much larger Dynamic of the UL Power Control

Furthermore in uplink the power control has much larger dynamics compared to downlink.
UL does not suffer from Channelization Code Shortage

Higher order modulation like 16-QAM alleviates the problem of scarce channelization codes available in downlink, which is not the case for uplink.
Fast Power Control cannot be abandoned in UL

Additionally the fast power control cannot be abandoned in the case of continuous uplink transmission due to the near-far problem. Simulation results have shown a significant improvement compared to Rel. 5, in the order of 50%-70% increase in system capacity, 20%-55% reduction in end-user packet call delay and around 50% increase in user packet call throughput, when simultaneously applying NodeB scheduling, hybrid ARQ with soft combining, and a shortened TTI. Hence, significant technical benefits have been found for a system using these techniques in conjunction. Complexity of hybrid ARQ has been studied in terms of buffering and timing requirements. The enhancements can be introduced into the FDD specifications without impacting the backwards compatibility with Rel. 5 and earlier releases.
1-56
MOTOROLA LTD.2002
Preview to HSUPA
Version 1 Rev 0
Preview to HSUPA
MOTOROLA LTD.2002
1-57
Version 1 Rev 0
HSUPA - Areas of Improvement

Based on research, the following enhancements for HSUPA are recommended:
NodeB controlled scheduling

With the physical layer scheduling in NodeB, the scheduling period could be shorter and the physical layer measurement information readily available in Node could be used as a basis for scheduling. This enables more up-to-date scheduling decisions and thus, better use of the available uplink air interface capacity. Significant reduction of the scheduling periods allows for more dynamic control over the uplink air interface capacity. Whenever one UE stops transmitting or reduces its transmission data rate, the released capacity can quickly and efficiently be allocated to other UEs.
Hybrid ARQ in Uplink

The principle of fast HARQ is to allow the NodeB to ask the UE to retransmit an uplink packet if it was received incorrectly. Additionally, the NodeB can use different methods for combining the multiple transmissions of a single packet, thus reducing the required Ec/No of individual transmissions. Therefore, the BLER target of the outer loop power control could be significantly set to higher values. A higher BLER target reduces the UE transmission power required for a given service/data rate. Hence, for the same cell loading, the cell capacity can be increased. How the uplink HARQ should operate when the UE is in soft handover is still under discussion, but a gain from HARQ in soft handover has already been proven. Note: HARQ operation in soft handover introduces one additional complexity not present in HSDPA HARQ. The gain from soft handover comes from one NodeB receiving a packet correctly, while another NodeB is failing in decoding. Hence, one NodeB sends a positive ACK to the UE and the other sends a NACK. In such a case, the network has received the packet and the UE should not send the same packet again. The SRNC needs to ensure the in-sequence delivery to higher layers and it needs to perform the selection combining for the packets received from different NodeBs.
Shorter TTI
Introducing a shorter TTI in the uplink reduces the HARQ retransmission delay. The motivation for the 2 ms TTI introduced in downlink HSDPA was due to the lack of power control being employed for the HS-PDSCH transmission. Such reasoning is not applicable in the uplink as disabling uplink power control would reduce the system capacity. The challenge for the 2 ms TTI stems from the resulting uplink range with limiting power resources in the UE. With an equal amount of data per TTI, it is possible to transmit less energy during 2 ms than e.g. compared to 10 ms. Additionally, the interleaving gain is reduced at 2 ms TTI. Thus, even if shorter TTIs were used due to the advantage for the HARQ, for the cell edge operation, the use of 10 ms TTI might be still be necessary in order to make the uplink more robust at low Ec/No conditions.
1-58
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-59
Version 1 Rev 0
HSUPA - Less Beneficial Areas

Higher Order Modulation
Higher order modulation, of which only 8-PSK has been studied, has been found to cause a loss in link performance compared to multi-code transmission with BPSK. In the uplink there is also no shortage of channelization codes like in the downlink as every UE has its own scrambling code and thus all UEs in the same cell can use the same channelization codes. As already known from EDGE, higher order modulation requires greater transmitter linearity and tighter constraints for error vector magnitude and thus makes the terminal hardware more complicated and also more expensive. Therefore it is plausible that research in higher order modulation types may not be further pursued.
Fast DCH Setup

Fast DCH setup has been partially investigated. Methods for reducing the synchronization time when going from CELL_FACH to CELL_DCH have been described in 3GTR 25.896 but not evaluated in detail. Other aspects of fast DCH setup, e.g. architectural changes and signaling protocols, have not been covered either until now but may be at a later stage.
1-60
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
1-61
Version 1 Rev 0
1-62
MOTOROLA LTD.2002
HSDPA Principles
Version 1 Rev 0
Chapter 2 HSDPA Principles
MOTOROLA LTD.2002
2-1
Version 1 Rev 0
HSDPA Principles
2-2
MOTOROLA LTD.2002
Objectives
Version 1 Rev 0
Objectives
After this Lecture the Student will be able to: Describe the coding chains of HS-DSCH, HS-SCCH and HS-DPCCH State the timing relations between DPCH and HS-DSCH and HS-DPCCH Describe the implications of using 16-QAM State the advantages and liabilities of AMC in HSDPA Describe difference between chase combining, partial IR and full IR
MOTOROLA LTD.2002
2-3
Version 1 Rev 0
HS-SCCH Frame Structure

The HS-SCCH subframe is shared between part 1 and part 2 in a 1:2 manner. Part 1, the TFRC, comprises the modulation type (QPSK or 16-QAM), channelization code set information and transport block size information. The TFRC is identified by the TFRI value which is only indirectly signaled on HS-SCCH through part 1 and 2. Part 2, the HARQ information, comprises the Hybrid-ARQ process information, redundancy and constellation version and new data indicator. Further details of the HS-SCCH structure are given below.
Part 1
Part 1 consists of the time critical HS-PDSCHs channelization code set and modulation scheme. Both parameters are vital for the physical decoding of HS-PDSCHs. This function is indicated by the arrow pointing from part 1 towards the HS-PDSCH subframe. In order to distinguish different UEs the H-RNTI is encoded into the part1 bits. The 16 bit value for H-RNTI is actually not present inside the part 1, but it is just masked with the part 1 bits. This method allows the layer 1 alone to determine if the HS-SCCH subframe contains a HS-DSCH resource assignment without layer 3 involvement. It can be assumed that invoking layer 3 for determining resource allocation would consume more time and processing. Therefore this method using physical layer is very efficient.
Part 2
Part 2 consists of several parameters. The H-RNTI is encoded together in part 2 and part 1 bits. Transport-block size information is used to derive the actual transport block size. The transport block size varies from 137 to 27,952 bits. Hybrid ARQ process information is signaled by RRC to maintain several transmission instances at the same time. Redundancy version: On each retransmission of a block, the set of bits that are actually transmitted may be changed for example, the set of parity bits chosen for transmission on the second attempt may be different to the set chosen on the first attempt. These different versions of the coding format for the data block are called redundancy versions. In order to attempt a decoding, the receiver needs to know the redundancy version used when each transmission attempt is made (in order to perform the HARQ combining process). Constellation version: only used in case of 16-QAM. An additional parameter which can be varied between retransmission of a block is the bit-to-constellation mapping (also referred to as Constellation Re-arrangement for details please refer to slide 16-QAM Constellation Rearrangement). New data indicator allows the UE to distinguish between initial transmission or retransmission within a HARQ process. Encoded UE Identity comprises the H-RNTI which is attached in the CRC checksum.
2-4
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-5
Version 1 Rev 0
HS-SCCH Coding Chain

RV Coding
b represents the constellation version parameter to swap the bit stream in case of 16-QAM. This averages the interference. s represents the indicator of self-decodable or non-self decodable transmission. r indicates the redundancy version for chase combining and IR packet re-transmissions. The output of the RV coder provides a 3 bit value which is called X(RV).
MUX of HS-SCCH Part 1

Channelization Code Set Info (7 bits) is mapped in the following manner: The OVSF codes shall be allocated in such a way that they are positioned in sequence in the code tree. That is for p multicodes at a starting offset y in the code tree the following codes are allocated: C(ch,16,y) C(ch,16,y+p-1). The number of multicodes p and the corresponding offset y to identify the HS-PDSCHs is signaled to the UE via the channelization code set info in the very HS-SCCH subframe. Modulation scheme determines if QPSK or 16-QAM is used for the subsequent HS-PDSCHs.
MUX of HS-SCCH Part 2

Following parameters are multiplexed together: New data indicator (1 bit), Redundancy and constellation version (3 bits), HARQ process info (3 bit), Transport block size info (6 bits) The 13 output bits of the multiplexer and the 8 bits from part 1 multiplexer are then CRC encoded together with the 16 bits obtained from H-RNTI. The 16 bits CRC check is then just appended to the 13 bits of part 2.
HS-SCCH Channel Coding and Rate Matching

The 29 bits of part 2 are piped through the 1/3 convolutional coder. In order to reset the flip-flops, 8 tail bits are added before coding. This encodes part 2 bits to 111 output bits. After rate matching, the number of encoded bits is reduced to 80 bits. The 8 bits of part 1 are also piped through the 1/3 convolution coder with 8 tail bits added. The encoded 48 bits are then punctured down to 40 bits
UE specific Masking of Part 1 bits

To speed up the detection of HS-DSCH resource allocation, the 16 bits of H-RNTI are 1/2 convolutional encoded which brings the amount of bits up to 48. After that the 48 bits are punctured down to 40 bits to obtain a unique 40 bits bit sequence used to mask the 40 bits of part 1.
2-6
MOTOROLA LTD.2002
Version 1 Rev 0
80 Part 1 Bits
H-RNTI Encoding
40 Part 2 Bits
MOTOROLA LTD.2002
2-7
Version 1 Rev 0
UE specific Masking for Part 1

The figure illustrates the CRC attachment for HS-SCCH. The post convolutional encoded part 1 bits are binary XORd with the H-RNTI. The H-RNTI was before convolutional coded and then punctured down to 40 bits in order to match the number of part 1 bits. The masking of part 1 bits with the H-RNTI is a so called scrambling where the part 1 bits are binary summed up with the H-RNTI bits through a modulo 2 operation which means XORing. Note: Due to the very high total transmission rate of the HS-DSCH (peak rate > 10 Mbit/s), it is impossible for a low complexity UE to buffer the entire HS-DSCH until successful receipt of the HS-SCCH and start decoding. Therefore it is very important that the UE knows, what modulation type and channelization codes belong to it in order to demodulate and do limited buffering only. Therefore this complexity limitation divides the HS-SCCH TTI into two parts, with part 1 consisting of the code and modulation information needing to be decoded before the arrival of the HS-DSCH. The de-scrambling, decoding and validation of part 1 information should be completed before the start of the corresponding HS-DSCH subframe transmission. As the HS-SCCH carries the signalling message for the HS-DSCH, clearly successful operation of the HS-DSCH relies on a low HS-SCCH error rate.
2-8
MOTOROLA LTD.2002
Version 1 Rev 0
H-RNTI Encoding
MOTOROLA LTD.2002
2-9
Version 1 Rev 0
UE Specific CRC Attachment for Part 1 and Part 2

The figure illustrates the CRC attachment for HS-SCCH. A 16 bit UE specific CRC checksum is computed over part 1 and part 2 and attached to part 2. This is achieved by simple XORing the CRC checksum from obtained from part 1 and part 2 with UEs H-RNTI. The output is then just appended to the sequence of part 2 bits. Note: At every HS-SCCH TTI, the UE monitors up to four HS-SCCHs and tries to extract part 1 information from them. Although there is a UE specific CRC at the end of part 2 to prevent falsely decoding for the unintended UE, the UE processing resource is wasted due to buffering of all four HS-SCCHs. This waste in resources is prevented by simply XORing the post convolutionally encoded part 1 with UEs H-RNTI.
2-10
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-11
Version 1 Rev 0
HARQ Process ID and New Data Indicator

The flow chart shows how the HS-DSCH transport blocks get evaluated per HARQ process. We would like to point out that the transmission and retransmission in HSDPA works entirely without using sequence numbers. There are important benefits when doing transmissions without sequence numbering. The HARQ information comprises of the HARQ process id and the NDI bit as well as the RV parameter. A NEW transmission of a HS-DSCH TB is signaled via HS-SCCH when the NDI bit toggles compared to the previous transmission. A RETRANSMISSION is signaled if the NDI bit remains the same. After decoding the UE sends via the HS-DPCCH either an Ack or Nack towards NodeB depending on the CRC check outcome. Note the important thing to understand here is that the transmission and potential retransmission is performed entirely without higher layer and thus does not need sequence numbering. This makes the transmission fast and robust. All the necessary information (TB size, HARQ process id and NDI bit) are signaled to the UE via HS-SCCH not on HS-DSCH. The fact that the HARQ info is per HARQ process introduces a simple stop and wait algorithm. The UE is hereby told to perform soft-combining of sub-sequent transmissions within the same HARQ process unless the NDI bit toggles.
HARQ Process ID
The part 2 of HS-SCCH informs the individual UE what HARQ process is currently being served. The HARQ process id identifies an ongoing TB transmission which remains active in both NodeB and UE until the TB has been successfully decoded by UE and the NodeB received an Ack for it. As long as the NDI bit (new data indicator) of a certain HARQ process does not toggle, every (re-) transmission is targeted for the same HARQ process. As Layer 1 soft-combining takes place per HARQ process, the UEs receiver is informed that a combined decoding out of original transmission and retransmission(s) should be performed. The fixed timing relation between downlink HS-DSCH TB transmission and sub-sequent uplink acknowledgment on HS-DPCCH makes the use of sequence numbering obsolete. The effect of this is twofold. Firstly no higher layer in UE needs to be involved to determine whether the sequence number is correct or incorrect which keeps the entire process very hardware oriented and therefore very fast (avoiding processing delays of higher layers). Secondly the sequence number does not even have to be transmitted reducing unnecessary overhead signaling which could be prone to errors too.
New Data Indicator

The NDI bit has to be considered per HARQ process. If continuous HS-DSCH transmissions are successful, the NDI bit toggles with every new HS-DSCH transmission. If the UE signals a Nack for a certain HARQ process, the NodeB will continue with retransmission(s) for that certain HARQ process. The number of retransmissions might be limited by the vendor (hard-coded in the software) and/or supervised by PDU release timer T1. Upon timer expiry retransmission are stopped and higher layer may take action. So retransmission of the same HS-DSCH data within the same HARQ process always have the same NDI bit value. This means for retransmission of the same HARQ process, the NDI bit does not toggle. Note after e.g. more than 3 retransmissions of the same data , the MAC-hs in NodeB may decide to flush the soft-buffer in UE and transmit the entire PDU fresh again. This is simply indicated to the receiver when the NDI bit is unchanged for a certain HARQ process but the transport block size is changed. The UEs receiver is then aware that a fresh transmission of the faulty TB is attempted by NodeB.
2-12
MOTOROLA LTD.2002
Version 1 Rev 0
les
MOTOROLA LTD.2002
2-13
Version 1 Rev 0
1-Channel Stop and Wait

The figure should highlight that HSDPA physical layer consists of state machines. They are so called working on a SAW (stop and wait) principle. The parameter HARQ process ID therefore represents a channel ID operating on SAW. When a subframe gets transmitted then NodeB has to wait for Ack/Nack before a new or (re-) transmission can be scheduled. In our case the NodeB can only transmit one subframe every 5 ms (depending on NodeBs processing delay and RTT, realistic is every 6-10 ms). The HARQ process in NodeB and UE remains in the state Nack as long as there the decoding remains unsuccessful. Once the UE reports an Ack, the HARQ process is successfully terminated but can be re-open again any time if the NodeB transmits for the same HARQ process ID a new subframe. In the figure we show how three HS-DSCH subframes get transmitted one after the other whereas the third subframe is a retransmission of the second subframe. In HSDPA the Layer 1 operates entirely without sequence numbers when transmitting the different HS-DSCH subframes to the UEs. Therefore at any moment in time, only one subframe per HARQ process can be distinguished as the Ack/Nack from the UE is also without sequence number. Actually the NodeB unambiguously relates the very Ack/Nack to the right downlink transmission as the Ack/Nack has to be transmitted within 5 ms after HS-DSCH subframe reception. So the fixed timing relations allows the transmission without sequence numbers. In order to supervise the correct reception of HS-DSCH subframes the NodeB needs to start a timer after allocating an HS-DSCH subframe. If the Ack/Nack from UE arrives in time, the timer gets stopped and either a new transmission ( UE transmitted an Ack) or a retransmission (UE transmitted a Nack) can be scheduled for the same HARQ process. In order to cope for wrong decoding of HS-SCCH at the UE and potential corruption of Ack/Nack in the uplink, the NodeB needs to have a watch-dog timer which expires slightly after 5 ms (actually the timer must consider the round trip time) if the neither Ack nor Nack has been received from UE. In order to have a continuous data flow to a UE, the 1-Channel Stop and Wait scheme must be enhanced to N-Channel SAW scheme. In HSDPA N is limited to 8 maximum 8 processes can be open by NodeB.
2-14
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-15
Version 1 Rev 0
HARQ Processes with N-Channel Stop and Wait

In order to reduce receiver buffering requirements a HARQ scheme based on N-channel stop-and-wait protocol has been specified. An example is depicted in the figure. UE-1 uses N-channel HARQ with N = 4 , UE-2 uses N-channel HARQ with N = 2 and 3 slot TTI.
Preconditions:
UE-1 is on 16-QAM and thus needs to obtain beside the RV also the CV for bit rearrangement (e.g. in case of retransmission). UE-1 might be very close to the NodeB and is exposed to a good C/I respectively high Ec/No in downlink. UE-2 has less favorable channel condition and is therefore on QPSK.
HS-SCCH Signaling
The HS-SCCH indicates the UE-identity and the modulation type used for HS-DSCH transmission in the subsequent HS-DSCH subframe. Although the HARQ process info, RV and CV and NDI (new data indicator) are signaled in the second part of HS-SCCH, we drew them directly above each HS-DSCH subframe. This should enhance the understanding of the picture. Note: NDI does not get incremented for retransmission..
HS-DSCH Transmission and Retransmission

Each HS-DSCH packet is ackd or nackd while other packets are transmitted so that the downlink channel can be kept occupied all the time if there are packets to transmit. The X(RV) for UE-1 indicates self-decodable transmissions and retransmission as s = 1. Thus partial IR is used for retransmission when TB #1 and TB #2 are retransmitted s = 1 & r = 1. Also the CV changes for retransmission b = 1. N-channel HARQ supports asynchronous transmission: different users can be scheduled freely without waiting for completion of a given transmission receiver needs to know which HARQ process the packet belongs to, which is explicitly signaled on HS-SCCH. The asynchronous feature of N-channel HARQ is also shown in the figure: after four packets to UE-1, a packet is transmitted to UE-2 and the retransmission to UE-1 is delayed by one TTI. The processing times should be such that continuous transmission to a UE is possible. In the uplink direction acknowledgements of the received MAC-hs PDUs are sent via HS-DPCCH. For UE-1, TB #1 and TB #2 are nackd and need to be re-sent. Due to strict timing requirements of N channel stop-and-wait HARQ scheme the processing time requirements both in UE and Node B should be carefully considered. At UE side there should be sufficient time left for all the processing after the reception of the HS-DSCH transport block of a TTI until the transmission of acknowledgement in uplink. Similarly, sufficient time should be left for Node B to react to the received acknowledgement message before the next TTI for the given HARQ process. The figure also shows an example of a feedback timing of a 4-channel stop-and-wait HARQ, e.g. considering process 0 of UE-1. While waiting for the acknowledgement, NodeB continuous to send three more HS-DSCH TBs to the same UE-1. With increasing the number (N) of parallel HARQ processes one will improve the time diversity per process but this results in longer available processing times both at UE and Node B. Downside of this are the increased buffering requirements at UE and longer delay per process. These factors should be taken into account when configuring the maximum for N via RRC. Note: N is limited to a maximum of eight.
2-16
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-17
Version 1 Rev 0
Multicode Operation in HSDPA

The figure shows two UEs operating in HSDPA.
HS-SCCH-set Decoding
The graphic demonstrates that both UEs have to decode their assigned HS-SCCH-set first, before they can attempt to decode the HS-PDSCHs. For simplicity reasons, both UEs have the same HS-SCCH-set assigned. A HS-SCCH-set consists of up to a maximum of four HS-SCCH channelization codes C(ch,128,1) C(ch,128,127). As depicted, only one of the four HS-SCCHs contains valid information per UE. This is indicated by the appropriate colour coding for each UE. All the information necessary for demodulating the related HS-PDSCH subframe which follows always 2 slots later is transmitted to the UEs within part 1 of each HS-SCCH. It can be seen that every HS-SCCH is logically divided into two parts. The second part contains the necessary information on how to decode the HS-DSCH. So part 1 and part 2 serve different purpose. Part 1 allows the demodulation of the HS-PDSCH subframe and part 2 is responsible for layer 2 decoding of the HS-DSCH. Note, that part 2 is fully covered in chapter 2 Note: The HS-SCCH is intended for the very UE once it recognizes its UE-id inside part1 of the HS-SCCH subframe. If the UE detected consistent control information intended for this UE in the immediately preceding subframe, it is sufficient to only monitor the same HS-SCCH used in the immediate succeeding subframe. In the graph this is indicated by e.g. user 1 in slot 0 to slot 6 where user 1 gets two consecutive valid HS-SCCHs. Therefore UE 1 only needs to decode the same HS-SCCH from slot 3 onwards.
HS-DSCH Demodulation
If a UE detects that one of the monitored HS-SCCHs contains its UE-id and consistent control information intended for this UE, the UE shall start receiving the HS-PDSCHs. Consistent control information hereby means that modulation scheme and HS-PDSCH channelization code-set info are valid according to the UEs capability. The UE has about one slot duration time after receiving part 1 to prepare for HS-PDSCHs reception. As already mentioned, the UE indicates via the category parameter if it supports up to 5, 10 or 15 HS-PDSCH channelization codes in parallel. The color coding used in the figure for the HS-SCCH and their related HS-DSCH shows that HSDPA allows for time multiplexing and code multiplexing of the HS-PDSCHs. Time multiplexing means that user 1 and user 2 get the HS-PDSCHs assigned one after the other in different subframes. Code multiplexing or multicode operation means that several user, here user 1 and user 2, use different HS-PDSCHs within the same subframe. The various HS-PDSCHs are separated by different channelization codes.
2-18
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-19
Version 1 Rev 0
HS-SCCH Power Control

The HS-SCCH power control is under the control of the NodeB. It may follow the power control commands sent by the UE to the NodeB via uplink DPCCH or any other power control procedure applied by NodeB. The HS-SCCH might additionally be power controlled by the CQI feedback. The HS-SCCH power should be unchanged during a TTI. Note: Although the HS-SCCH is power controlled, its extremely short frame size (2 ms) provides for little time diversity. With presence of fading and channel quality feedback delay, it may turn out that the power required to guarantee e.g. 1% BLER is quite large. Since the HS-SCCH is sharing power with the HS-DSCH and other dedicated channels, this incurs a significant loss in the system capacity. The power consumption and margin of the HS-SCCH can be reduced by considering transmit and / or receive diversity. Considering the additional channel condition feedback required for the closed loop schemes and the robustness under variable mobile speeds, open-loop transmit diversity schemes might be preferred over the closed-loop diversity schemes.
2-20
MOTOROLA LTD.2002
Version 1 Rev 0
2-21
MOTOROLA LTD.2002
Version 1 Rev 0
HS-PDSCH Frame Structure

The subframe and slot structure of HS-PDSCH are shown in the figure. The HS-PDSCH is used to carry the HS-DSCH. The related subframe of HS-PDSCH is delayed by 2 x 2560 chips compared to the corresponding HS-SCCH subframe. A HS-PDSCH corresponds to one channelization code of fixed spreading factor SF=16 from the set of channelization codes reserved for HS-DSCH transmission. Multi-code transmission is allowed, which translates to a UE being assigned multiple channelization codes in the same HS-PDSCH subframe, depending on its UE category. NOTE Remember a HSDPA capable UE either supports up to 5, 10 or 15 physical channels of HS-PDSCH in parallel.
A HS-PDSCH is transmitted either with QPSK (mandatory to be supported) or 16-QAM modulation. The symbol rate of HS-PDSCH is obtained by following calculation: 38400 chips / (16 chips / symbol) = 2400 symbols / (10 ms) = 240 ksps. Within a HS-DSCH subframe all the allocated HS-PDSCHs per user must use the same modulation type, e.g. 16-QAM. However for a different user in the same HS-DSCH subframe the remaining HS-PDSCHs can be transmitted using a different modulation type, e.g. QPSK. This is code multiplexing. The potential drawback is that the processing requirement (and delay) for the NodeB becomes quite high when doing extensive multi code operation. In QPSK, each symbol represents 2 bits, whereas in 16-QAM each symbol corresponds to 4 bits. The maximum gross data rate with the maximum number of channelization codes for HS-PDSCH can be calculated like this: QPSK: 480 kbit/s x 15 physical channels = 7.2 Mbit/s 16-QAM: 960 kbit/s x 15 physical channels = 14.400 Mbit/s Taking the maximum transport block size which can be signaled via HS-SCCH HSDPA: Practical maximum transport channel throughput rate: 27952 bits / 2 ms = 13.976 Mbit/s NOTE All relevant signalling information is transmitted in the associated HS-SCCH i.e. the HS-PDSCH does not carry any signalling information on how to decode the HS-DSCH TB.
2-22
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-23
Version 1 Rev 0
HS-PDSCH Channelization Code Tree

The figure gives an example how the available channelization codes under the primary scrambling code may be shared between HSDPA and Rel. 99 or Rel. 4 users. In HSDPA, every HS-PDSCH uses a fixed spreading factor of SF = 16 Up to 15 codes in parallel may be allocated to a UE. The OVSF channelization code tree is allocated by the C-RNC. The HSDPA codes are autonomously managed by NodeBs MAC-hs scheduler function.
Example:
The picture shows 8 consecutive codes reserved for HS-DSCH, starting at C(ch,16,8) and ending with C(ch16,15). Additionally, HS-SCCH codes with SF = 128 need to be allocated for the HS-SCCH-set. The HS-SCCH-set can comprise up to 4 HS-SCCHs thus giving the C-RNC a wider flexibility in sharing the OVSF tree between Rel. 5 and legacy release UEs. The HS-DSCH channelization-code-set information signalled over HS-SCCH is mapped as follows: The OVSF codes shall be allocated in such a way that they are positioned in sequence in the code tree. That is, for p multicodes at offset y the following codes are allocated: C(ch,16,y) C(ch, 16,y+p-1) For the example presented here, the parameter y and p have to be set like this: y = 8 and p = 8.
2-24
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-25
Version 1 Rev 0
HSDPA Code- and Time Multiplex Operation

The picture shows an example of how the channelization codes available for HSDPA can be shared in the time and code domain. There are 6 subframes depicted for HS-DSCH and 7 for the HS-SCCH-set. Example: 5 codes are reserved for HSDPA, 1 or 2 UEs are active within one TTI
Resource sharing in code as well as well time domain

In case of multi-code transmission, a UE is assigned to multiple codes in the same TTI. (e.g subframe 0). Multiple UEs may be assigned channelization codes in the same TTI (e.g subframes 14).
UE procedure for receiving HS-DSCH

When the UE monitors HS-SCCHs, the UE shall only consider the control information to be consistent if decoded channelization-code-set information is lower than or equal to maximum number of HS-DSCH codes received according to its UE capability and if the decoded modulation scheme is valid according to its UE capability. If the UE did detect consistent control information intended for this UE in the immediately preceding subframe, it is sufficient to only monitor the same HS-SCCH used in the immediately preceding subframe. That shall be indicated by subframe 2 to 5in the figure where UE #1 and UE #3 receive further HS-DSCH allocations in the consecutive subframe. If a UE detects that one of the monitored HS-SCCHs carries consistent control information intended for itself, the UE shall start receiving the HS-PDSCHs indicated by this control information. The transport block size information shall be derived from the signaled part 2 value as defined explained in slide HS-SCCH Frame Structure. If the Hybrid-ARQ process information is not included in the set configured by upper layers, the UE shall discard the information received on this HS-SCCH and on the HS-PDSCHs.
2-26
MOTOROLA LTD.2002
Version 1 Rev 0
Subframe 0
Subframe 1
Subframe 2
Subframe 3
Subframe 4
Subframe 5
Subframe 6
Subframe 0
Subframe 1
Subframe 2
Subframe 3
Subframe 4
Subframe 5
MOTOROLA LTD.2002
2-27
Version 1 Rev 0
Uplink HS-DPCCH Frame Structure

The figure illustrates the frame structure of the HS-DPCCH. The HS-DPCCH carries uplink feedback signaling related to downlink HS-DSCH transmission. The HS-DSCH-related feedback signaling consists of HARQ-ACK (Hybrid-ARQ Acknowledgement) and CQI (Channel Quality Indication). Each subframe of length 2 ms (3 x 2560 chips) consists of 3 slots, each of length 2560 chips. The HARQ-ACK is carried in the first slot of the HS-DPCCH sub-frame. The CQI is carried in the second and third slot of every HS-DPCCH subframe. There is at most one HS-DPCCH per HSDPA-UE. The HS-DPCCH can only exist together with an uplink DPCCH. The timing of the HS-DPCCH relative to the uplink DPCCH is shown latter in this chapter. The spreading factor of the HS-DPCCH is 256. The 10 bit signaling data per uplink HS-DPCCH slot is encoded to 30 bits for error protection. There exists only one slot format for uplink HS-DPCCH In the uplink direction the HS-DPCCH has to carry Ack/Nack and quality feedback information. Ack/Nack info indicates to the NodeB if the transmission was successful or if retransmission has to be scheduled. The quality feedback info allows the NodeB to determine at what data rate new transmissions and possible retransmission can be sent to UE. One requirement is that a soft handover still works regardless of whether all NodeBs involved support HSDPA or not. Thus, it was decided to leave the existing uplink channel structure unchanged and add the needed new information elements for HSDPA on a parallel code channel, namely HS-DPCCH.
HARQ-Ack
The Ack/Nack transmission reflects the results of the CRC check after the packet decoding and also possible combining with former retransmission.
Channel Quality Indication

CQI indicates which estimated transport block size, modulation type and number of parallel codes could be received correctly within reasonable BLER in the downlink direction.
2-28
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-29
Version 1 Rev 0
Uplink HS-DPCCH Code Allocation

For the DPCCH, DPDCHs and HS-DPCCH the following applies: The DPCCH is always spread by channelization code C(c) = C(ch,256,0). The HS-DPCCH is spread by channelization code C(hs) = C(ch,256,hs) written in the presented table. The channelization value hs depends on the number of active DPDCHs.
Multicode Operation in Uplink

When only one DPDCH is to be transmitted, DPDCH(1) is spread by code C(d,1) = C(ch,SF,k) where SF is the spreading factor of DPDCH(1) and k= SF / 4. When more than one DPDCH is to be transmitted, all DPDCHs have spreading factors equal to 4: DPDCH(1) C(ch,4,1) DPDCH(2) C(ch,4,1) DPDCH(3) C(ch,4,3) DPDCH(4) C(ch,4,3 DPDCH(5) C(ch,4,2) DPDCH(6) C(ch,4,2) PDCH(n) is spread by the channelization code C(d,n) = C(ch,4,k), where k = 1 if n {1, 2}, k = 3 if n {3, 4}, and k = 2 if n {5, 6}.
2-30
MOTOROLA LTD.2002
Version 1 Rev 0
Code (ch, 256, hs)
MOTOROLA LTD.2002
2-31
Version 1 Rev 0
Uplink HS-DPCCH Coding Chain

HARQ-Ack and CQI data are channel coded and then mapped onto the physical channel. Both indicator HARQ-Ack and CQI are processed in parallel as the flows are not directly multiplexed but are transmitted at different times onto HS-DPCCH. The HARQ-Ack is transmitted onto the first slot followed by CQI which is transmitted onto the subsequent two slots.
Channel Coding for HS-DPCCH

Two forms of channel coding are used, one for the channel quality information (CQI) and another for HARQ-ACK. Channel Coding for HARQ-Ack The HARQ-Ack message (1 bit) to be transmitted, is coded to 10 bits as shown in the presented table. In case of Ack, 10 times repetition of 1 is coded and in case of Nack, 10 times repetition of 0 is coded. Channel Coding for CQI The CQI is coded using a (20,5) code. The code words of the (20,5) code are a linear combination of the 5 basis sequences Mi,n with n = 1..5. The meaning of (20,5) is that the code codes 5 input bits to 20 code-bits, thus 5 data bits are channel coded to 20 code-word bits. This corresponds to a coding rate of 5/20 = 1/4. This is a lot of redundancy and therefore provides for a quite good error protection. The CQI values 0 .. 30 are converted from decimal to binary to map them to the channel quality information bits (1 0 0 0 0) to (1 1 1 1 1) respectively. The information bit pattern (0 0 0 0 0) shall not be used in Rel. 5. The CQI bits are a(0), a(1), a(2), a(3), a(4), where a(0) is LSB and a(4) is MSB. Note: For simplicity reasons the 5 sequences for Mi,n are not shown. Example: Tabulated CQI value = 14 1binary conversion = 01110. a(0) = 0, a(1) = 1, a(2) = 1, a(3) = 1, a(4) = 0 Calculation of i(0) i(19): i(0) = (0*1 + i*0 + 1*0 + 1*0 + 0*1) mod 2 = 0 i(1) = (0*0 + 1*1 + 1*0 + 1*0 + 0*1) mod 2 = 1 i(2) = (0*1 + 1*1 + 1*0 + 1*0 + 0*1) mod 2 = 1 ... i(19) = (0*0 + 1*0 + 1*0 + 1*0 + 0*1) mod 2 = 0
2-32
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-33
Version 1 Rev 0
Uplink HS-DPCCH Spreading

The Figure illustrates the principle of the uplink spreading of DPCCH, DPCHs and HS-DPCCH.
Spreading
Spreading is applied to the physical channels. It consists of two operations. The first is the channelization operation, which transforms every data symbol into a number of chips, thus increasing the bandwidth of the signal. The number of chips per data symbol is called the spreading factor. The second operation is the scrambling operation, where a scrambling code is applied to the spread signal.
DPCCH / DPDCH / HS-DPCCH

The binary DPCCH, DPCHs and HS-DPCCH to be spread are represented by real-valued sequences, i.e. the binary value "0" is mapped to the real value +1, the binary value "1" is mapped to the real value 1. The value "DTX" which is only applicable for HS-DPCCH is mapped to the real value 0. The DPCCH is spread to the chip rate by the channelization code 256,0. The n-th DPDCH called DPDCH(n) is spread to the chip rate by the channelization code C(d,n). The spreading factor ranges hereby from 4 to 256. The HS-DPCCH is spread to the chip rate by the channelization code C(256,hs). Depending on the number of DPDCHs, HS-DPCCH is mapped onto the I-branch in case of even number of DPDCHs. If there is an odd number of DPDCHs, then HS-DPCCH is mapped onto the Q-branch. Note: More than one DPDCH is only allowed for SF 4. If only one DPDCH is transmitted, then DPDCH(1) is spread by C(d,1) = C(ch,SF,k) where SF is the spreading factor of DPDCH(1) and k = SF/4. One DPCCH, up to six parallel DPDCHs, and one HS-DPCCH can be transmitted simultaneously 1 n 6.
Gain Factors
After channelization, the real-valued spread signals are weighted by gain factors, (c) for DPCCH, 2(d) for all DPDCHs and 2(hs) for HS-DPCCH (if one is active). Note: A maximum of one CCTrCH is allowed for one UE on the Uplink. It can be either one CCTrCH of dedicated type or common type.
2-34
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-35
Version 1 Rev 0
Uplink Transmission Delay T(0)

This picture demonstrates why a 1024 chip delay is introduced between downlink and uplink transmission.
Uplink / Downlink Timing at UE

At the UE, the uplink DPCCH/DPDCH frame transmission takes place approximately T(0) chips after the reception of the first detected path (in time) of the corresponding downlink DPCCH/DPDCH frame. T(0) is a constant defined to be 1024 chips.
Uplink Fast Power Control Reason for T(0)

The figure shows the reason for the 1024 chips delay(1024 x 0.26 x 10E-6 sec = 266.24 micro seconds). T(0) is vital for the fast inner loop power control achieved by TPC bits of DPCCHs. 1. In uplink the NodeB measures in slot N-1 the SIR of the UEs pilot channel and calculates a SIR(estimate). The SIR(estimate) is compared with the SIR(target) given by the outer loop power control. Note: The RNC is in charge of updating the SIR(target) based on transport block CRC check. If SIR(estimate) < SIR(target), the NodeB commands the UE to power up. If the SIR(estimate) > SIR(target) the NodeB commands the UE to power down. In any case, the power up or down command is transmitted to the UE via TPC in slot N still arriving at UE in slot N-1 due to the 1024 chips delay compared to the downlink. The UE processes the TPC and updates the output power of the DPCH (DPDCH and DPCCH) accordingly in slot N. Thus starting with the pilot field, the new power value has been already set by UE in slot N.
2. 3.
The downlink fast power control works similarly.
2-36
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-37
Version 1 Rev 0
HS-SCCH / HS-DSCH and HS-DPCCH Timing (1)

The figure should highlight the timing relations between Rel. 99 DL DPCH and the determination of the uplink timing of the HS-DPCCH. In order to be 100% orthogonal to the UL Rel. 99 DPCCH, the UE need to derive the factor m from the received DL DPCH and the difference between DL DPCH and start of transmission of the relevant HS-DSCH subframe. The timing value T(TX_diff) represents the time difference between the DL Rel. 99 DPCH and the start of the respective HS-DSCH subframe. As the duration of DPCH is 10 ms, there are always 5 subframes corresponding to the DL Rel. 99 DPCH. Therefore T(TX_diff) consists of 5 different timing values which always differ by 3 slots 7680 chips. The figures shows a special case where the timing of Rel. 99 DL DPCH t(DPCH) coincides with the transmission of HS-DSCH subframe 0. For this special case the T(TX_diff) = 0 chips For T(TX_diff) = 0 and considering that above timing is measured at UE (DL DPCH timing is 1024 chips advanced compared to UL DPCCH): Timing at UE: 101 x 256 chips + 1024 chips = 19200 chips + 7680 chips The above equation proves that for the uplink the constant of 101 x 256 chips + 1024 chips always corresponds to the downlink timing of 19200 chips + 7680 chips. For any of the HS-DSCH subframes the T(TX_diff) always ensures that the UE transmits in time and orthogonal to the uplink DPCCH by a multiple of 256 chips.
2-38
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-39
Version 1 Rev 0

The figure shows the timing offset between the downlink DPCH, uplink DPCH, the HS-PDSCH and the uplink HS-DPCCH at the UE. Note that the transmission delay over the air-interface in downlink direction is not considered here. (The DL DPCH arrives with zero delay at the UE.)
Uplink HS-DPCCH
A HS-DPCCH sub-frame starts m x 256 chips after the start of an UL DPCH frame. The UL DPCH corresponds to the DL DPCH frame from the HS-DSCH serving cell. The value m is calculated as: m = (T(TX_diff)/256 ) + 101
T(TX_diff)
T(TX_diff) is the difference in chips (T(TX_diff) = 0, 256, ....., 38144), between a) the transmit timing of the start of the related HS-PDSCH subframe (see HSDPA Downlink Timing in chapter 1) b) the transmit timing of the start of the downlink DPCH frame from the HS-DSCH serving cell that contains the beginning of the HS-PDSCH subframe Therefore T(TX_diff) is the resulting time difference between firstly the time relation of HS-SCCH and DL DPCH and secondly the time relation between HS-SCCH and the beginning of the HS-PDSCH subframe. Note: At any one time, m therefore takes one of a set of five possible values according to the transmission timing of HS-DSCH sub-frame timings relative to the DPCH frame boundary. The UE and Node B shall only update the set of values of m in connection to UTRAN reconfiguration of downlink timing. Example calculation for the second HS-DSCH subframe and its related uplink HS-DPCCH: T(TX_diff) = [t(start of HS-DSCH-subframe) t(DPCH,n) versus start of HS-SCCH)] T(TX_diff) = [(5120+7680) 2560] = 10240 m = (T(TX_diff) / 256) + 101 m = (10240 / 256) + 101= 141 x 256 chips = 36096 chips Rule:101 x 256 chips + 1024 chips + T(TX_diff) = 19200 chips + 7680 chips+ T(TX_diff) Note: The multiple of 256 chips offset between HS-DPCCH and uplink DPCCH ensures orthogonality between the two physical channels.
2-40
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-41
Version 1 Rev 0
Delta Ack / Nack and Ack-Nack Repetition Factor

DeltaAck
Integer value ranging from (0..8). Determines power increase of HS-DPCCH slot carrying Ack. The power increase is relative to the associated uplink DPCCH
DeltaNack
Integer value ranging from (0..8) Determines power increase of HS-DPCCH slot carrying Nack. The power increase is relative to the associated uplink DPCCH.
ACK-NACK-repetitionFactor
Integer value ranging from (1..4) The UE shall transmit the Ack/Nack information received from UEs MAC-hs in the slot allocated to the HARQ-ACK in the corresponding HS-DPCCH sub-frame as shown in slide HS-SCCH / HS-DSCH and HS-DPCCH Timing. When ACK-NACK-repetitionFactor is greater than one, the UE shall: repeat the transmission of the Ack/Nack information over the next (ACK-NACK-repetitionFactor-1) consecutive HS-DPCCH sub-frames, in the slots allocated to the HARQ-ACK and not attempt to receive nor decode transport blocks from the HS-PDSCH in HS-DSCH sub-frames corresponding to HS-DPCCH sub-frames in which the Ack/Nack information transmission is repeated. If consistent control information is not detected (i.e. UE is not able to decode its H-RNTI) in the downlink HS-SCCH subframe on any of the HS-SCCHs in the HS-SCCH set, neither Ack, nor Nack shall be transmitted in the corresponding HS-DPCCH subframe.
2-42
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-43
Version 1 Rev 0
Delta Ack / Nack and CQI Power

When an HS-DPCCH is active, the power offset 3(HS-DPCCH) for each HS-DPCCH slot shall be set as follows. For HS-DPCCH slots carrying HARQ Acknowledgement: 3(HS-DPCCH) = 3(ACK) if the corresponding HARQ Acknowledgement is equal to 1 3(HS-DPCCH = 3(NACK) if the corresponding HARQ Acknowledgement is equal to 0 For HS-DPCCH slots carrying CQI : 3(HS-DPCCH) = 3(CQI) Note: The values for 3(ACK), 3(NACK) and 3(CQI) are set by RRC signaling In non-compressed frames (hs) is the gain factor which is calculated according to
where 4(c) value is signaled by higher-layer, e.g. through RRC or calculated based on TFCI selection. With the exception of the start and end of compressed frames, any DPCCH power change shall not modify the power ratio between the DPCCH and the HS-DPCCH.
2-44
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-45
Version 1 Rev 0
HS-DPCCH ACK / NACK and DTX Recognition

The table shows an estimated likelihood of erroneous reception of HS-DPCCH Ack/Nack at the NodeB. In the first four cases, the NodeB allocates HS-DSCH subframes via HS-SCCH part 1 and part 2. We show what potential errors can occur in the reception of HS-SCCH and HS-DSCH at UE and the HS-DPCCH reception at NodeB. The last two cases should give the reader an idea that it cannot completely ruled out that the UE determines a successful resource allocation in part 1 and/or part 2 although the NodeB did not transmit on HS-SCCH to that particular UE. Such false alarms can be decoded by the UE either completely out of the noise (part 1 and part2 ) or due to coincidence that part 1 bits match a proper resource allocation by chance after decoding. The logic used for ACK/NACK/DTX signaling by the UE is as follows: An ACK (logical 1 bit) is signaled if the UE-specific CRC in part 2 of one of the monitored HS-SCCHs passes the check implying that this UE considered itself scheduled and the data-specific CRC check passes on the associated HS-PDSCH implying that the packet was successfully decoded requiring no further retransmission. A NACK ( logical 0 bit) is signaled if the UE-specific CRC in part 2 of one of the monitored HS-SCCHs passes the check implying that this UE considered itself scheduled and the data-specific CRC fails on the associated HS-PDSCH implying that the packet was not yet successfully decoded requiring further retransmission. Neither Ack nor Nack is signaled in any other case. The UE shuts off or so called DTXes the ACK/NACK field. This happens when the UE-specific CRC in part 2 of all monitored HS-SCCHs failed the check or such check was not attempted due to the poor metric quality obtained from decoding the UE ID scrambled part 1 information of the HS-SCCH implying that this UE considered itself not scheduled on any of the HS-PDSCHs. This could also happen when the part 2 CRC check passes for more than one of the monitored HS-SCCHs that is normally impossible, hence suggesting an irresolvable error situation.
False Alarm
In very rare cases another UE decodes successful part 1 and part 2 and transmits either Ack/Nack in the uplink although no resources were allocated to this UE at all via HS-SCCH. So the NodeB should ignore that Ack/Nack as it knows that no resource allocation was sent to that UE via HS-SCCH. Note: Every UE has its own uplink scrambling code thus ACK/NACK and/or CQI repetitions can overlap among different UEs in uplink. Both procedures are independent from each and therefore may work asynchronous to each other (not every CQI repetition has a preceding ACK/NACK repetition).
2-46
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-47
Version 1 Rev 0
CQI Description and Task

TFRC Info
The CQI value consists of a recommended TFRC provided by the UE to Node B. When a UE reports a particular CQI, it is reporting that, for the current radio conditions, the UE is able to receive data with a transport format corresponding to the reported CQI, and lower CQIs, at single-transmission BLER probability no greater than 0.1, taking the total received HS-PDSCH power of P(HSPDSCH) = P(CPICH)+ 5 + 3 (in dB) into account. If S-CPICH is used as reference, the power offset is with respect to the S-CPICH used by the UE; otherwise the P-CPICH is the reference. The measurement power offset 5 is signaled to the UE using RRC signaling. The reference power adjustment 3 is defined in reference tables. TFRC reporting as opposed to C/I reporting is used in order to account for the disparity between more and less advanced receivers that for the same observed channel conditions support higher or lower rates. The CQI for different UE capabilities is determined from TFRC reference tables. For each entry in the table, there is an associated turbo-encoder information word size (expressed as TBS plus 24-bit CRC), modulation type, and number of HSPDSCH codes. These parameters describe the reference, single-transmission modulation, and coding scheme associated with each CQI value.
Out of Range Signaling

Note that if - for the current radio conditions and measurement power offset 5 - the UE cannot support the minimum CQI (CQI = 1) at BLER probability <= 0.1, then the UE reports CQI = 0, which indicates an out of range condition to the Node B.
Power Reduction Factor 1

To ensure that no UE indicates a transport format exceeding its capabilities, an HS-PDSCH power reduction factor value 3 (also known as reference power adjustment) is used to indicate radio conditions beyond the highest transport format. The HS-PDSCH power reduction factor 3 should be interpreted such that the UE is able to receive data with the highest supported transport format assuming P(HS-PDSCH) = P(CPICH) + 5 + 3
Measurement Power Offset 2

5 indicates an positive or negative offset of the HS-PDSCH power compared to the received CPICH power.
CQI Reporting Cycle

The frequency of the CQI reports as well as the sub-frame offset during which UE transmits CQI is controlled by RRC signaling and HS-DSCH transmission. Lower frequency cycles are intended for use for certain (or all) UEs when they have low activity and/or relatively static channel and/or low capability and/or whenever uplink load on the Node B has to be regulated.
2-48
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-49
Version 1 Rev 0
HS-DPCCH Measurement Feed Back - CQI

The depicted physical layer parameters are signaled to UE and NodeB from RRC.
Measurement Power Offset 3

The measurement power offset, 6 is the power offset to be applied on the measured CPICH power. P(HSPDSCH) = P(CPICH) + 6+ 3 Range for. 6: (-6 .. 13 by step of 0.5) dB.
CQI Feedback cycle, k

1) UE derives the CQI value as defined in clause CQI Mapping Table UE Category 2) For k = 0, the UE shall not transmit the CQI value. 3) For k > 0, the UE shall transmit the CQI value in each subframe that starts m 256 chips after the start of the associated uplink DPCCH frame with m fulfilling: (5 x CFN + [m x 256 chips/7680 chips ]) mod k with k = k / (2 ms) where CFN denotes the connection frame number for the associated DPCH and the set of five possible values of m is calculated as described in sub clause HS-SCCH / HS-DSCH and HS-DPCCH Timing 4) The UE shall repeat the transmission of the CQI value derived in 1) over the next (CQI-RepetitionFactor 1) consecutive HS-DPCCH sub frames in the slots respectively allocated to the CQI as defined in sub clause HS-SCCH / HS-DSCH and HS-DPCCH Timing. Note: UE does not support the case of k < CQI-RepetitionFactor. The UE shall not transmit the CQI in other subframes than those described in 3) and 4). Range for feedback cycle k: Integer (0, 2, 4, 8, 10, 20, 40, 80, 160) unit: ms
CQI-RepetitionFactor
Integer value ranging from (14) Determines how often the CQI is repeatedly sent from UE towards NodeB to indicate the downlink radio channel quality.
DeltaCQI
Integer value ranging from (08). Determines power increase of HS-DPCCH slot carrying CQI. The power increase is relative to the associated uplink DPCCH.
2-50
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-51
Version 1 Rev 0
CQI Reporting Principle

The CQI estimation and reporting is indeed a quite difficult task for the UE. As can be seen on the left side in the figure, a iterative algorithm might be implemented in UE to derive the CQI value. The motivation of CQI reporting is to give the MAC-hs in NodeB an indication about the actual channel transfer capacity. The MAC-hs in NodeB may take the CQI value into account for deriving the TFRC for UEs next HS-DSCH transmission. Therefore MAC-hs in NodeB is informed about: Based on the last HS-DSCH data transfer and the current DL channel quality a new data transfer with transport block size X, number of HS-PDSCHs and modulation Z would be possible. Hereby it is guaranteed, that the BLER probability is less or equal 10%.
UE measures the downlink channel quality

UE may measure the downlink SIR of the received HS-PDSCHs if a HS-DSCH was previously allocated. UE measures the received quality (Ec/No) of the CPICH in a 3 slot reference period ending one slot before the start of the first slot in which the CQI value is transmitted. UE estimates the CQI value through an iterative process taking the measurements of HS-PDSCH_SIR and CPICH_Ec/No into account. Beside this the UE probes with varying parameter values for TB size, modulation type, a reference power setting for HS-PDSCH and an possible power adjustment value to obtain a CQI value. The result considering the used modulation type, transport channel block size etc. shall sustain a BLER <= 10%.
2-52
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-53
Version 1 Rev 0

CQI Algorithm in UE
The estimation of the CQI value is a difficult task for the UE and has to be verified i.e. through simulations. If the UE reports a certain CQI value depending on its category, i.e. CQI = 7, then this value and all lower values (1 to 6) give the NodeB an indication what transport block sizes X, number of HS-PDSCHs and modulation type Z could sustain a BLER probability of less or equal 10%. Of course the NodeB may vary the parameters comprising TFRC depending on the current resource status. The updown arrows indicate which input parameters are tunable in order to reach a BLER <= 10%. The reference power adjustment 3 indicates a possible reduction of the HS-PDSCH transmission power as the downlink channel quality exceeds the possible range and physical capability of the UE.
2-54
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-55
Version 1 Rev 0

NodeB TFRC decision
The NodeB determines the TFRC (TB size, modulation type and number of spreading codes) based on UEs CQI value which must be seen as a reference only and might be not applied 1:1 in downlink resource assignments. Possibly also DL power control commands of the associated DPCH might be taken into account however when DPCH is in soft handoff they become quite inaccurate. Also the Ack/Nack ratio gives an indication in the long run about the DL reception quality and also if the chosen TFRCs were too optimistic or too conservative. However, the exact algorithm in NodeB is vendor dependent. Note: The UE has limited time for CQI reporting (e.g. 5 ms) especially when there was a previous HS-DSCH reception or if the CQI feedback-cycle is set to 2 ms.
2-56
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-57
Version 1 Rev 0
CQI Mapping Table UE Category

UE Category
The UE category together with the supported modulation type determine which column to select. The Parameters for UEs supporting only QPSK can be found in category 11 and 12.
CQI value
The CQI value 0 is reserved for the state do not bother to transmit while the other states represent the transmission capacity that the UE can receive at the current time. The CQI value consists of 5 bits that carry quality related information.
Transport Block Size

The TBS indicates in bits per 2 ms subframe the amount of data the UE is currently able to digest. It considers the UE capability maximum number of HS-DSCH transport channel bits that can be received within one HS-DSCH TTI.
Reference Power Adjustment

For the purpose of CQI reporting, the UE shall assume a total received HS-PDSCH power of P(HSPDSCH) = P(CPICH) + 6 + 3 in dB. The total received power is evenly distributed among the HS-PDSCH codes of the reported CQI value, the measurement power offset 6 is signaled by higher layers and the reference power adjustment 3 is shown in both tables here.
N(IR) and X(RV)

The UE shall assume the number of soft channel bits available in the virtual IR buffer N(IR) as well the RV (redundancy and constellation version) parameter X(RV) depending on its category. Note the X(RV) is a result of further parameter s, r and in case of 16-QAM b. All of them are explained in full detail in slide HS-SCCH Coding Chain.
2-58
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-59
Version 1 Rev 0
Example for a CQI Report

Here we would like to give you an idea how the NodeB shall interpret the CQI values. We chose a UE with Category 6 which reports CQI value = 15. TB size = 3319 bits total number of bits = (3319 + 24 CRC) x 3 + 12 = 10041 bit For Full and Partial IR there exists the potential problem of exceeding the UEs indicated soft-buffer of 9600 bits. 441 parity bits need be punctured or Chase Combining must be employed, as it does not exceed the UEs IR buffer of 9600 bits with QPSK. The UE calculated the CQI for Chase Combining and an IR buffer of 9600 bits for retransmission as the BLER probability is less or equal 10% thus retransmission needs to be considered. The standard method for CQI calculation considering retransmission is Chase Combining as X(RX) = 0 ( s = 1, r = 0, b = 0) If the NodeB decides to use 16-QAM the maximum number of bits which can be transferred using 5 HS-PDSCHs = 9600 bits. Using Chase Combining for potential retransmission there is no problem in exceeding UEs soft-buffer of 9600: R 3319 / 9600 If Full or Partial IR is/are used for retransmission, then the turbo coded bits of 10041 bits need to be reduced to 9600 bits by puncturing the parity bits only.
2-60
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-61
Version 1 Rev 0
CQI and Ack/Nack Considerations

HS-DPCCH Power Control during Soft Handover
The HS-DPCCH uses certain power offsets (signaled by CRNC through RRC to UE) that are applied to the Ack, Nack or CQI transmissions and these offsets are defined with respect to the power of the UL-DPCCH pilot (Rel. 99 / Rel. 4). Therefore the HS-DPCCH received signal power is also power controlled in accordance with the Rel99 / R-4 inner loop power control. This, however, can cause problems when the UE is in soft handoff and receiving multiple power control commands at the same time from several NodeBs. In the typical situation of OR of DOWNs, ( UE shall power down when at least one radio link indicates power down) the result is that the UL-DPCCH pilot and the HS-DPCCH received signal power at the HS-DSCH serving Node B (for whom the HS-DPCCH information is destined) can be faded well below normal levels even at low vehicle speeds, in which case the adverse effect is longer lasting. Furthermore, channel estimation based on the UL-DPCCH pilot in order to de-rotate and detect the HS-DPCCH symbols suffers badly. Ensuring that the long-term error rate requirements on the CQI and ACK/NACK channels are met would then require quite large power offsets. This can have a significant impact on the coverage range of the UE (when its transmit power is limited) and on the uplink capacity (due to the injection of a large amount of interference to the HS-DSCH serving Node B as well as surrounding Node Bs).
CQI Repetition
As a partial solution to the problem outlined above, it is recommended to have UTRAN-controlled repetition factors for the Ack/Nack and CQI fields in order to decrease the power offsets. The repetition factors for these two fields are set independently of each other and range from 1 through 5. The repetition takes place across consecutive sub-frames taking the value of m and the value of k into account. For more details please see page HS-DPCCH Measurement Feed Back CQI. Note: When the ACK/NACK repetition factor is greater than one, then the UE will not be scheduled for any new or retransmitted data in the corresponding successive HS-DSCH TTIs.
Intra NodeB Softer HO

In case of Intra NodeB cell change, both cells in the active set of the Rel. 99 DPCH may receive and decode the HS-DPCCH. As both signals terminate in the same NodeB, a maximum ratio combining might be feasible. This would improve the successful decoding of the HS-DPCCH and therefore reduce layer 1 retransmission and thus also RLC retransmission. Note: When the ACK/NACK repetition factor is greater than one, then the UE will not be scheduled for any new or retransmitted data in the corresponding successive HS-DSCH TTIs.
2-62
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-63
Version 1 Rev 0
CQI Reporting Feedback Cycle and Repetition

The figure should show how CQI repetition and feedback cycle have to be interpreted as well as Ack/Nack repetition factor setting. The UE transmits the ACK/NACK information received from MAC-hs in the slot allocated to the HARQ-ACK in the corresponding HS-DPCCH sub-frame as defined ~ 5 ms after HS-DSCH subframe reception.
Ack/Nack Repetition Factor

When ACK/Nack repetition factor is greater than one, the UE shall: repeat the transmission of the ACK/NACK information over the next ack_nack_repetition_factor 1 in consecutive HS-DPCCH sub-frames, in the slots allocated to the HARQ-ACK and not attempt to receive nor decode transport blocks from the HS-PDSCH in HS-DSCH sub-frames corresponding to HS-DPCCH sub-frames in which the ACK/NACK information transmission is repeated.If consistent control information is not detected on any of the HS-SCCHs in the HS-SCCH set, neither ACK, nor NACK, is transmitted in the corresponding HS-DPCCH subframe. CQI Reporting Cycle with k = 2 ms or k = 4 ms The picture should enhance the fact that CQI reporting and Ack/Nack reporting do not necessarily occur within the same subframe. UE-1 follows a feedback cycle of 4 ms. This means every 4 ms the UE-1 transmits a CQI value. As the CQI repetition-factor is set to 2,a new report is sent only after two CQI transmissions UE-2 follows a feedback cycle of 2 ms. This means every 2 ms the UE-2 transmits a CQI value. As the CQI repetition-factor is set to 4, every fourth report only contains a new calculated value.
Trade off between Ack/Nack/CQI Repetition and long Feedback Cycles

If Ack/Nack repetition factor is greater than 1, then the NodeB cannot allocate HS-DSCHs in continuous subframes. Thus such a setting is only recommended if the UE capability does not support continuous subframe allocation or there is a only a low to medium throughput required. If the UE does not have to served within every subframe, the CQI reporting cycle can be set to a longer value, like in our picture for UE-1 to 4 ms. The setting of repetition factors for CQI and Ack/Nack greater than 1 brings two advantages to the UTRAN. Firstly the probability of successful decoding at the NodeB is increased and secondly the gain setting may be reduced.
2-64
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-65
Version 1 Rev 0
HSDPA during Compressed Mode Operation

The figure shows HSDPA operation during DPCH compressed mode. For simplicity reasons, the transmission gap is the same for uplink and downlink and a double gap is depicted. Compressed mode is always activated by SRNC when the associated DPCH of HSDPA carries an AMR speech call and an event 2d (bad FDD coverage) was reported by UE. During compressed mode on the associated DPCH, the following applies for the UE for transmission of HS-DPCCH and reception of HS-SCCH and HS-PDSCH: 1. The UE neglects a HS-SCCH or HS-PDSCH transmission, if a part of the HS-SCCH or a part of the corresponding HS-PDSCH overlaps with a downlink transmission gap on the associated DPCH. In this case, neither ACK, nor NACK is transmitted by the UE to respond to the corresponding downlink transmission. 2. If a part of a HS-DPCCH slot allocated for ACK/NACK information overlaps with an uplink transmission gap on the associated DPCH, the UE does not transmit ACK/NACK information in that slot. 3. If in a HS-DPCCH sub-frame a part of the slots allocated for CQI information overlaps with an uplink transmission gap on the associated DPCH, the UE does not transmit CQI information in that sub-frame. 4. If a CQI report is scheduled in the current CQI field (as shown on page 103) and the corresponding 3-slot reference period (as shown on page 95) wholly or partly overlaps a downlink transmission gap, then the UE should use DTX in the current CQI field and in the CQI fields in the next (N_cqi_transmit 1) subframes. Summary: DPCH compressed mode has priority over HSDPA operation.
2-66
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-67
Version 1 Rev 0
Effects of Using 16-QAM

Higher Throughput Rates
Compared to QPSK, 16-QAM doubles the throughput rate, thus exploiting the available resources more efficiently than QPSK. Obtained throughput with a code rate of and 15 channelization codes in parallel: QPSK: Channel Bit Rate = 480 kbit/s 15 x 480 kbit/s x = 5.4 Mbit/s 16-QAM: Channel Bit Rate = 960 kbit/s 15 x 960 kbit/s x = 10.8 Mbit/s
Increased Spectrum Efficiency

16-QAM allows the transmission of more bits/Hz compared to QPSK. HSDPA with 16-QAM makes use of the property of high Ec/No available in a cell, usually close to the NodeB. QPSK is limited in the throughput rate as it can not increase the throughput anymore in areas with strong C/I. However with 16-QAM and simultaneously utilizing of up to 15 codes in parallel the entire cell capacity might be useable.
Higher Interference Vulnerability

16-QAM is more sensitive to interference compared to QPSK at the air-interface as both the phase and the amplitude are modulated. On top of this modulation inaccuracy will lead to higher retransmission.
Smaller Decision Space

When looking at the I/Q diagram the phase separation compared to QPSK has decreased with 16-QAM to about 1/3. The channel estimation in 16-QAM requires amplitude and phase to separate the constellation points on the I/Q diagram. Due to amplitude modulation, the transmitter linearity requirement increases compared to QPSK. Note: The relative power ratio between CPICH and HS-DSCH needs to be estimated in 16-QAM thus no fast power control is allowed with subframes.
2-68
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-69
Version 1 Rev 0
16-QAM Constellation Rearrangement

Note: This function only applies to 16-QAM. In case of QPSK it is transparent. The purpose of re-arranging the constellation of the symbols in the I/Q plane is to average the level of error probability after retransmission combining. When 16-QAM is used, two of the four bits constructing the received symbols will have higher probability of error than the other two. In order to compensate for this effect, it is possible to use constellation re-arrangement for retransmission which provides for a swapping of the bit streams. Using standard Gray labeling, the first two bits of the symbol have higher reliability than the second two bits. The positions i1 and q2 having higher reliability and are marked in the figure (left hand side) The constellation rearrangement can take one out of four possible values. The figure shows the first two values. The initial setting for 16-QAM is b = 0. The bit rearrangement parameter b = 1 provides for a swapping of i1 with i2 and q1 and q2.
2-70
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-71
Version 1 Rev 0
16-QAM Constellation Rearrangement for b = 2 and b = 3

Bit Re-arrangement b = 2
Output bit sequence = i1q1i2q2 XOR with 0011 which is equivalent to inversion of logical values of i2 and q2
Bit Re-arrangement b = 3
Output bit sequence = i2q2i1q1 Swapping i1 with i2 and q1 with q2 and XOR with 0011
2-72
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-73
Version 1 Rev 0
Power Sharing between HSDPA and DCH

Node B Transmit Power
The picture shows an example of power resource sharing between non-HSDPA users (e.g. speech traffic only) and HSDPA users on the same carrier. The power allocated to HSDPA may be dynamically adjusted by the CRNC according to the instantaneous traffic requirements. HSDPA uses AMC (also known as fast link adaptation) and does not need spare room for fast power control.
Maximum Transmission Power

The Maximum Transmission Power is the maximum value for the linear sum of the power of all downlink physical channels, that is allowed to be used in a cell. If transmit diversity is applied to one downlink physical channel, the power to be considered for this downlink physical channel is the linear sum of the power used for this downlink physical channel on all branches. The reference point is the antenna connector. Furthermore, with HSDPA the DCH is typically used to carry time critical signaling information and conversational type scenarios, such as speech in parallel with HSDPA operation. Thus user can benefit from a simultaneous connection on the DCH in parallel with for example a high speed download via HS-DSCH. Note: The NodeB has to assure with appropriate power control setting for a secure decoding of the HS-SCCH. As the HS-SCCH cannot be retransmitted, it is therefore vital that the UE safely decodes part 1 and part 2 of HS-SCCH. Otherwise the subsequent HS-DSCH cannot be decoded.
UE Transmit Power
The HS-DPCCH needs some part of the uplink transmission power, which has an impact on the link budget for uplink. With active HS-DPCCH the uplink power increases in the network resulting in a higher uplink interference value. At the same time the UEs battery capacity decreases. Note: On the HS-DPCCH there is some possibility to set an optimum between deltaAck/Nack and the ack-Nack-repetition-factor. The same is possible between deltaCQI and CQI-repetition-factor. A higher repetition factor requires less power increase for CQI and Ack/Nack reporting. However there is a trade off between consecutive HS-DSCH allocation (also depends on the UE category) and minimizing uplink interference. As long as the Ack/Nack is repeated, no new HS-DSCH can be allocated. A high repetition factor for CQI reporting may delay further HS-DSCH allocation as well as AMC in NodeB is not able to decode the reported CQI value as the UE transmitted with too low uplink transmission power. So there is some room for optimizing the system performance.
2-74
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-75
Version 1 Rev 0
Adaptive Modulation and Coding

Main Benefits of AMC
The advantages of adapting transmission parameters in a wireless system with respect to changing radio conditions have been formerly approved in e.g. EDGE. In general, the process of modifying transmission parameters to adapt to the current channel conditions at the air interface is known as LA (Link Adaptation). The benefits of AMC are: Availability of higher throughput and therefore increased average throughput per cell. The modulation type and code rate can vary from subframe to subframe. Reduced interference variation due to modulation and coding based link adaptation instead of variations in transmit power. The code rate and modulation type can be adjusted according to CQI and DL transmit power value of DCH as well on the Ack/Nack ratio. Utilization of short term fading in the way that a user might be always served in a constructive fade. The relative fast CQI reporting allows a very quick change in the modulation type and code rate thus making use of 16-QAM and highest code rate whenever possible. This maximizes user and cell throughput and exploits good and extreme good C/I conditions of the individual UEs. Boosted effectiveness in combination with time and code multiplexing scheduling techniques. With HSDPA UEs can receive nomore than one physical channel per subframe. Also more than one UE can be served in one TTI as the available channelization codes can be shared among several UEs according to their capability. This is a significant enhancement compared to Rel. 99 where multicode operation is only possible with SF=4.
2-76
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-77
Version 1 Rev 0
Adaptive Modulation and Coding - Principle

The picture shows the syntactical application flow of AMC as proposed for HSDPA from NodeBs point of view. The principle of AMC is to enable the system to change the modulation and coding format. Therefore, the channel condition has to be measured or estimated on feedback of the receiver. In this way, links with better radio transmission conditions are assigned a higher order modulation and higher coding rates, while links with bad radio transmission conditions are assigned lower order modulation and lower code rates.
AMC Scheduling Function

The AMC scheduler is a key element of the design in HSDPA. It controls the allocation of the channel to users, and to a large extent, it determines the overall behaviour of the system. The scheduler exploits the multi-user diversity and statistical gain and therefore strives to transmit to users when radio conditions are favourable, permitting high data rates. Nevertheless, it also maintains a certain degree of fairness. Fundamentally, higher tolerance to QoS criteria means higher system capacity. The outcome of this is one of the main enhancements for best effort service, which by definition allows for a relatively large spread in quality. As the scheduling algorithm is not standardized, we provide at a later stage several different schedulers.
Layer 1 Processing Chain

The picture shows how the HS-DSCH is processed by layer 1 in NodeB The transport block is coded by a CRC checksum which allows for backward error detection in the receiver The coded bits are then turbo encoded by r = 1/3 meaning the tripling of the bit-rate HARQ provides for LA and variable code rate QPSK / 16-QAM stage routes the bit stream according to the designated modulation scheme Interleaving ensures that bit errors are distributed in a random fashion preventing that consecutive bits are affected by interference DEMUX de-multiplexes the HS-DSCH CCTrCH onto one or more physical channels The physical channels are spread by the respective channelization codes of spreading factor 16 The individual physical channels are complex added and fed into the modulator (QPSK or 16-QAM)
2-78
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-79
Version 1 Rev 0
Physical Layer Processing Chain of AMC

Methods of selecting the suitable MCS
CQI: The UE estimates the downlink channel quality and calculates the best fitting transport format and reports it to the serving NodeB MAC-hs estimates the best fitting transport format based on power control gain of the associated DPCH with or without feedback from UE MAC-hs takes the Ack/Nack information into account MAC-hs considers the queue priority and queue length of individual users based on higher layer signaling (RRC layer) Depending on the system load and channel conditions, MAC-hs adapts modulation and coding rate for each user.
AMC Processing Chain

CRC Attachment: Error detection is provided on HS-DSCH TB through 24 bit CRC. CRC calculation is same as for Rel. 99 channels. Bit Scrambling: The output bits are scrambled avoiding special repeated patterns e.g. zero patterns. Such a repeated pattern can cause some bias in Tx power of 16-QAM transmission which will lead to an inaccurate estimation of amplitude reference in 16-QAM demodulation. Code Block Segmentation: If the number of bits in a TTI (2 ms) exceeds 5114, being the maximum size of a code block in HSDPA, then code block segmentation is performed. This assures that the turbo coder still works efficiently. Channel coding for HS-DSCH shall be done with the general method as in Rel. 99 using following specific parameters. There will be maximum of one TB and the rate 1/3 turbo coding shall be used. Hybrid-ARQ functionality matches the number of bits at the output of the channel coder to the total number of bits of the HS-PDSCH set to which the HS-DSCH is mapped. HARQ is controlled by the RV (redundancy version) parameters which determine whether the encoded TB shall be self-decodable or non-self-decodable. The exact set of bits at the output of the HARQ functionality depends on the number input bits, number of physical channel capacity / UE soft buffer capacity and the RV parameters. Note: The HS-DSCH hybrid ARQ functionality is fully explained in chapter 3. Physical Channel Segmentation: When more than one HS-PDSCH is used, the encoded bits are divided among the different physical channels. HS-DSCH Interleaving: The interleaving is done separately for each physical channel. Note: The HS-DSCH Interleaving is fully explained in chapter 3 Constellation Re-arrangement for 16-QAM: This function is transparent for QPSK. The 4 bits per 16-QAM symbol mapping is re-arranged according to constellation version parameter b. Physical Channel Mapping: The input bits of the physical channel mapper are mapped in one subframe for one HS-PDSCH. The bits are mapped to the PhCHs (physical channels) so that the bits for each PhCHs are transmitted over the air in ascending order.
2-80
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-81
Version 1 Rev 0
AMC Liabilities
AMC Liabilities
The implementation of AMC in UMTS also offers several challenges.
UEs measurement inaccuracy

The most critical point of AMC is its sensitivity to error and delay of channel quality measurements. To select the suitable transport format for current radio transmission conditions, the scheduler must always be aware of the current channel quality. Errors in channel estimation either cause selection of the wrong data rate and transmission at too high power, wasting system capacity or transmission with an MCS too sensitive towards bad channel conditions, increasing BLER.
Delay in CQI Reporting

Delay in reporting CQI measurement reduces reliability of the channel quality estimation due to its constant variation. The fast fading of the Uu requires a very fast tracking of the channel quality. However the scheduling frequency of HS-DPCCH is limited. Furthermore changes in the interference add to the measurement errors.
Limited Granularity for Data Rate Selection

AMC provides limited granularity in data rate selection and estimates of link quality are prone to error because of delay between the time of measurement and the time of rate selection. Unfortunately, AMC isnt enough to ensure an error free reception at UE. Therefore HARQ is needed for HSDPA since Link adaptation via AMC is insufficient for efficient retransmission of packet data. HARQ allows for a flexible code rate and retransmission methods using either chase combining or incremental redundancy.
NodeB Processing and Scheduling Delay

There is a natural transmission delay until a CQI report has been completely received in NodeB (2 slots plus uplink transmission delay). After that it depends on NodeBs processing speed to determine the appropriate transmission parameters (TFRI and HARQ parameters) based on the suggested CQI value. A further delay is added by the packet scheduler which has to consider the priority of the (re-) transmission based on vendor dependent algorithm. Finally NodeB is able to schedule the very HS-DSCH transmission and needs to derive the proper power values for HS-SCCH and HS-PDSCHs. Note: HARQ enables the implementation of AMC by reducing the number of required MCS levels and the sensitivity to measurement error and traffic fluctuations.
2-82
MOTOROLA LTD.2002
AMC Liabilities
Version 1 Rev 0
AMC Liabilities
MOTOROLA LTD.2002
2-83
Version 1 Rev 0
HSDPA HARQ Types
HSDPA HARQ Types

Hybrid Automatic Repeat reQuest (HARQ) is a protocol for error control in packet data transmission. When the receiver detects an error in a packet, it automatically requests the transmitter to resend the packet. This process is repeated until the packet is error free or the error continues beyond a predetermined number of transmissions and finally higher layers stop the retransmission. Possibly the erroneous packet is discarded. The users application is then in charge of dealing with packet losses. HARQ provides robustness through fast retransmission at the physical layer controlled by NodeBs MAC-hs. Retransmitted copies are combined at the receiver and then decoding is attempted again. There are three types of HARQ in HSDPA:
Chase Combining
Retransmission(s) of the same packet as that of the first attempt occur. The decoder combines multiple received copies of the coded packet weighted by their SNR prior to decoding. This method provides time diversity gain and is very simple to implement. Time diversity gain is simply the fact that the fast fading and interference changes between first transmission and multiple retransmission thus not all (re-)transmissions are affected by the same (bad) radio channel conditions.
Partial IR
The retransmission takes place with a partially different packet from the first one. Each packet transmitted in the partial IR scheme is self-decodable because it has the systematic bits of turbo codes. Instead of sending simple repeats of the entire coded packet, additional redundant information is incrementally transmitted. Note: The systematic bits of turbo encoded bits shall not be punctured, the other bits (parity bits p1 and parity bits p2) may be punctured.
Full IR
The retransmission of a fully different packet from the first one occur. The retransmission of packets are not self-decodable, thus they may contain only the parity bits of the turbo code output. This means also that the first transmission of a packet must at least contain the systematic bits of a packet. IR usually yields better performance compared to chase combining. However, it requires more implementation complexity and may not result in good performance unless the link adaptation errors are very small. Chase combining yields reasonable performance with lower implementation complexity and cost. Note: HARQ is controlled by MAC-hs scheduler.
2-84
MOTOROLA LTD.2002
HSDPA HARQ Types
Version 1 Rev 0
HSDPA HARQ Types
MOTOROLA LTD.2002
2-85
Version 1 Rev 0
HARQ Operation
HARQ Operation
Tthe following sections describe the principle of HARQ for RLC-AM. All three figures (chase combining, partial IR and full IR) always show at the end the status of the demodulator. A tick means that the demodulator could reliably predict whether the received bits are determined as logical 0 or 1. This information is used as input for the turbo decoder which decodes the HS-DSCH TB. Only the CRC check after turbo decoding ensures if the HS-DSCH TB was successfully received.In these sections, the aim is to demonstrate the differences in transmission and retransmission by using chase combining, partial and full IR The receiver comprises a demodulator and a channel decoder. The demodulator generates soft decision bits for each received bit. The soft decision bits represent the likelihood of the real bit value. Thus the demodulator tries to find out all possible combinations of bit values for unknown bits targeting the maximum probability of the bit sequence. The demodulator stores, therefore after each packet has been demodulated, the highest probability values for each bit in soft decision bits. When the channel decoder performs the decoding and the CRC check indicates a block error, retransmission is requested from the sender. The HS-DSCH is turbo encoded with a rate of 1/3. The turbo coder delivers 3 output bit streams. Systematic bits Parity P1 bits Parity P2 bits The systematic bits are always needed to decode the original transport block. Transmission or retransmission containing the systematic bits are therefore also called self-decodable transmissions. The parity bits add redundancy to the encoded block and are not sufficient to decode the transport block without systematic bits.
Chase Combining Performance

The retransmission of the same packet occurs always with the same puncturing scheme. Every transmission and retransmission contains systematic bits. The same puncturing scheme is indicated by the same two question marks in each packet. Thus chase combining makes use of time diversity. Not every transmission or re-transmission is affected by the same interference or fading. After the first instance of a packet has been received the demodulator keeps the soft decision values in a buffer as the CRC check indicated block error and a second instance of the same packet is requested. Once the second instance has been received, the demodulator can add up the soft decision values of each transmission based on their SNR value and so achieve a more reliable demodulation result. By adding up the soft decision values after every instance of a retransmission, the buffer capacity for the soft decision bits is modest compared to IR. Note: In chase combining, multiple transmissions of the same data packet are always performed with the same puncturing scheme. Every transmission is self-decodable as it contains the systematic bits.
2-86
MOTOROLA LTD.2002
HARQ Operation
Version 1 Rev 0
HARQ Operation
MOTOROLA LTD.2002
2-87
Version 1 Rev 0
HARQ Operation
HARQ Operation
Operation of Partial IR
With partial IR, every transmission and retransmission is still self-decodable however the puncturing scheme varies between retransmission. Thus the retransmission becomes more effective compared to chase combing where the puncturing scheme stays the same. As multiple retransmissions vary in their punctured bits, the demodulator can benefit from several instances of the same packet reducing the code rate. After each retransmission the demodulator can re-calculate the maximum likelihood by either taking the soft decision values of the transmission and retransmission into account or their added sum weighted by their individual SNRs. The buffer requirements are still modest compared to full IR Note: IR is another Hybrid ARQ technique wherein instead of sending simple repeats of the entire coded packet; additional redundant information is incrementally transmitted if the decoding fails on the first attempt. Partial IR is called Hybrid ARQ type-III as each retransmission is restricted to be self-decodable. In the IR schemes the receiver buffers coded symbols, which introduce new information to the HSDPA TTI transmitted first. Therefore the amount of data to be buffered increases with consecutive retransmission. Prior to decoding these symbols are soft-valued, i.e. each symbol is represented by two or more bits.
2-88
MOTOROLA LTD.2002
HARQ Operation
Version 1 Rev 0
HARQ Operation
MOTOROLA LTD.2002
2-89
Version 1 Rev 0
HARQ Operation
HARQ Operation
Operation of Full IR
Using full IR, the first transmission of a packet must be self-decodable thus the systematic bits need to be inside. However with every retransmission instance of the same packet only the parity bits may be transmitted. Every retransmission therefore reduces the code rate and increases the decoding probability. However the likelihood calculation makes it necessary to buffer each transmission instance. The soft decision values cannot be added up as the retransmissions containing the parity bits only are not self decodable. The full IR requires the biggest memory size and thus may not be useable for maximum data rate transmissions. Note: Full IR is called Hybrid ARQ type II as each retransmission is not self-decodable. In case of HARQ type II or type III with multiple redundancy versions, additional redundancy bits are sent during each retransmission yielding potentially more coding gain than simple type III with single redundancy version.
2-90
MOTOROLA LTD.2002
HARQ Operation
Version 1 Rev 0
HARQ Operation
MOTOROLA LTD.2002
2-91
Version 1 Rev 0
Scheduling Strategies Examples

Note: On the next slide we present several different schedulers: Round Robin, Proportional Fair and MAX-C/I.
AMC Scheduling Function

Basically the AMC scheduler function could base its decision on: The predicted or estimated channel quality The NodeB receives from the UE the CQI of the current downlink channel quality together with the Ack/Nack information. The Ack/Nack description tells the NodeB if a transport block was received correctly or not. On top of this the NodeB may use also the DL transmit power information on associated DCH to estimate the downlink quality. The downlink quality then determines the used modulation type and code rate for HS-DSCH transmissions. The current load of the cell The cell load may decide the type of PS (packet scheduler function) to be used. If only a few users are in the cell, a PS could be used to maximize users throughput. When the number of users rises, then a fairer PS might be advantages to serve each individual user at least according to its guaranteed throughput value and/or jitter requirement of its application. The traffic priority class (e.g. streaming class or background class) The traffic class could be used to determine how much delay is tolerable for a users application. User with background traffic class can be served less compared to user with streaming traffic class with need a more or less constant throughput. QoS AMC has to take into account the current QoS of a user (e.g. guaranteed throughput, guaranteed delay) compared to the requested or necessary QoS to support the particular traffic type.
2-92
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-93
Version 1 Rev 0
Comparison of Basic Packet Scheduler Methods

Scheduling Rate
The period over which users are scheduled ahead in time. If short, the packet scheduler may utilize fast channel variations and track fast fading for low-mobility users. Shorter periods call for higher computational complexity in the NodeB.
Serve Order
The order in which users are served; e.g. random order (round robin) or according to channel quality (C/I or throughput based). More advanced mechanisms require higher computational processing in NodeB.
Allocation Method
The criterion for allocating resources, e.g. same data amount or same power / code / time resources for all queued users per allocation interval (e.g. TTI). FT (Fair Throughput) Allocation method: Resources are allocated according to equal data amount for all users up to a maximum allocation. The users are served in random order. In theory, all users currently active will therefore experience the same delay and throughput. P-FR (Proportional Fair Resources) Allocation method: The same resources (power / code / time) are allocated for all users. Due to the fast scheduling frequency, the P-FR PS is able to track the fast fading. P-FR utilizes the CQI as well as link quality estimation algorithms, which are located in the NodeB. The proportional scheduling method results in all users getting approximately an equal probability of becoming served even though they may experience very different channel quality M-C/I (Maximum C/I) Allocation method: The same resources (power / code / time) are allocated per TTI. Like P-FR, the M-C/I scheduling rate is fast and therefore it is able to track the fast fading). The M-C/I scheduling method is the most drastic one, which serves the best user during the current TTI. For example, the user who can sustain the highest throughput. An available option with the HSDPA concept is to make very fast scheduling, which tracks the fast fading variations. Ultimately, users are only scheduled when they are experiencing constructive fading. This improves both, the user throughput and cell throughput for time-shared channels. However if QoS requirements dominate the scheduling strategy, the differentiation between the above-mentioned PS strategies become less significant as the PS strives to fulfill the users requirements instead of treating either everyone fair or special if when considering the channel quality. Event the capacity of the most aggressive schedulers, e.g. M-C/I reduces when QoS must be ensured. Actually they become more fair.
2-94
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-95
Version 1 Rev 0
Packet Scheduler in NodeB

Note: 3GPP does not specify the algorithm of the PS (packet scheduler) in NodeB for HSDPA. It is therefore up to the UTRAN vendors and Operator to make up their minds about suitable scheduling algorithms to satisfy their clients.
Fairness
The key to HSDPA operation and performance lies in the PS entity. Fast PS methods track the user channel quality for each TTI and makes scheduling decisions on the fly. The important issue is that of inherent fairness among users. The FT (fair throughput) approach ensures that all simultaneously queued users receive the same average throughput, which means that users in bad conditions are given relatively more HS-DSCH resources. On the other end, the M-CI (maximum C/I) method only gives resources to the users in the most favorable conditions thereby enhancing cell throughput at the expense of inter-user fairness. In between, the FR (fair resource method gives equal resources to all users so that all users will achieve a throughput corresponding to the channel quality.
Opportunistic Scheduling
The proportional methods P-FT (proportional fair throughput) and P-FR (proportional fair resource) attempt to give all users the same probability of being scheduled by using a relative instantaneous channel quality; e.g. defined as the ratio between current and average user Es/No. Ideally, this method will schedule users only during constructive fades thereby raising both the overall cell throughput and the user data rates. By weighting the relative instantaneous channel quality with buffer statistics, the fairness of the proportional methods can be increased (e.g. P-FT type scheduling).
Minimum Throughput Guarantees

In order to satisfy streaming services the NodeB is informed about e.g. the MAC-hs Guaranteed Bit Rate IE indicating the guaranteed number of bits per second that MAC-hs should deliver over the air interface under normal operating conditions (provided there is data to deliver).
Effect of TCP/IP
Avoiding delay jitter is essential to achieve high throughput for TCP. In particular, delay spikes can cause spurious timeouts. Such timeouts force TCP into slow-starts, which reduce TCPs congestion window sizes drastically. For HS-DSCH, jitter can occur due to varying interference as well properties of the PS influence the jitter.
Influence of Coverage
As mentioned already in chapter 1, there is no soft handover possible on the HS-DSCH. Therefore users at the cell edge do not benefit from soft handover gain when transmitting on the HS-DSCH. In order to transmit to these users at satisfactorily throughput levels, HS-DSCH must be transmitted with relatively higher power compared to users closer to the center of the cell. As the overhead associated with control signaling on HS-SCCH and HS-DPCCH is also highest at cell edge, it may be desirable to limit the coverage of the HS-DSCH. Note that the cost in carried load can be very high when supporting the full cell area as the effective throughput declines with increasing coverage. So less revenue per effective throughput can be achieved at large cell sizes.
2-96
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-97
Version 1 Rev 0
Packet Scheduling Strategies

The scheduler has to allocate the HS-DSCH resources to all HSDPA users in a more or less fair manner and try at the same time to reach a high utilization of the HSDPA resources. A high effective cell throughput (taking retransmission into account) is desirable to gain maximum revenue per transmitted byte. However, both strategies exclude themselves mutually. Note: Interactive, Background and Streaming traffic classes may be supported in HSDPA. Therefore the PS is will become one of the most complicated functionalities in HSDPA when trying to manage all three traffic classes at once, e.g. for the same user.
Fair Throughput Scheduling

Users get the same throughput no matter where they are. They are served in random order. In theory all users currently active in the system will therefore experience the same delay and throughput.
Fair Time Scheduling

The available power for HS-DSCH is shared equally among the users which lead to lower bit rates at the cell border. Users receive equal resources in random order and thus experience different data rates according to their average channel quality.
Long Term C/I Scheduling

It maximizes on the one hand the spectral efficiency but is very unfair as it only serves the best user during the current TTI, e.g. the user who can sustain the highest throughput
Proportional Fair Resource Scheduling

In order to obtain a fairer scheduling method, it is possible to define and calculate a so called RICQ (relative instantaneous channel quality) measure as a selection and prioritization metric. The RICQ utilizes the CQI information as well as the link quality estimation algorithms (e.g. DL power control) which are located at the NodeB. This fast scheduling method is referred to as the P-FR (proportional fair resource) scheduling method. The P-FR scheduling method results in all users getting approximately an equal probability of becoming active even though they may experience very different average channel quality. The principle of a packet scheduler: A PS must optimize the cell capacity while fulfilling the QoS requirements determined by a certain QoS policy. For Streaming Traffic, the QoS attributes to be satisfied are the Transfer Delay and the Guaranteed Bit Rate. There are no QoS guarantees for an Interactive Bearer, so the actual bearer quality in terms of minimum throughput will depend on the load of the system and the admission control policy of the operator. Background Traffic can be always transmitted when there is spare capacity in the network.
2-98
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-99
Version 1 Rev 0
Multi-User Selection Diversity Dynamic Scheduling

The PS is located in the NodeB. This opens up for fast scheduling at every TTI (2 ms) in HSDPA. The scheduling algorithm may be optimized according to the current channel conditions ( RICQ) which can yield large gain at cell and user levels. Basic diversity mechanism utilized is multi-user diversity = selection diversity among the active users
Dynamic Scheduling according to best C/I

Dynamic scheduling at the NodeB provides optimized usage of radio resources and exploits the short-term variations on the radio channels. It allows for a certain degree of QoS. There is a closed loop due to the physical feedback on HS-DPCCH about reception quality and Ack/Nack. There is a statistical multiplexing of data packets from different data flows (here from UEs) over one shared medium, namely HS-DSCH. Optimized usage of radio resources Exploitation of the short-term variations on the radio channels Provides certain degree of QoS
Scheduler Resources include:

Transmission power for HS-SCCHs and HS-PDSCHs Channelization codes for HS-PDSCHs Transport block size must be one out of a predefined set or at least compatible with UE category Number and codes for HS-SCCHs NodeB hardware/Processing boards configured for HS-PDSCHs and HS-SCCHs NodeB hardware/Processing boards configured for reception of HS-DPCCH per user individually
MAC-hs Scheduler should not allocate any HSDPA Resources to a User which:
Has no data for transmission Has no free capacity in any reordering buffer Has no ready-to-transmit HARQ process Has not completed the minimum inter-TTI interval after the last HSDPA transmission (see chapter 3 page HSDPA Category and Reference Combinations) Has reported a too low/bad CQI or CQI is out of range Has not reported any CQI or is physically out of sync in UL
2-100
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
2-101
Version 1 Rev 0
2-102
MOTOROLA LTD.2002
Forward and Backward Error Correction
Version 1 Rev 0
Chapter 3 Forward and Backward Error Correction
MOTOROLA LTD.2002
3-1
Version 1 Rev 0
Forward and Backward Error Correction
3-2
MOTOROLA LTD.2002
Objectives
Version 1 Rev 0
Objectives
Describe the Turbo Coding and Decoding Principle State the reason for Hybrid ARQ in HSDPA Describe the HARQ Functionality State the different purpose of Rate Matching Describe different performance of Partial IR and Full IR
MOTOROLA LTD.2002
3-3
Version 1 Rev 0
Turbo Coding
Turbo Coding
Turbo Coder Principle
In the beginning of this chapter we would like to introduce only the basics of turbo coding and decoding. This is necessary to ease the understanding of the Hybrid ARQ functionality which follows later on. First of all the fundamental turbo code encoder is built using two identical RSCs (recursive systematic convolutional coder) with parallel concatenation. An RSC encoder is typically r = 1/2 and is termed a component encoder. The two component encoders are separated by an interleaver. Only one of the systematic outputs from the two component encoders is used, because the systematic output from the other component encoder is just a permuted version of the chosen systematic output. The figure shows the fundamental turbo code encoder with an r = 1/3. One input bit corresponds to three output bits N (Sys, P_1 and P_2). The first RSC encoder outputs the systematic bits denoted as Sys and recursive convolutional sequence of parity bits denoted as P_1. The second RSC encoder discards its systematic bit sequence and only outputs its recursive convolutional sequence of parity bits denoted as P_2. Both encoder components are parallel concatenated and the same input data bits are fed in, therefore these codes are also called PCCC (parallel concatenated convolutional code).
RSC Encoder
The RSC encoder is obtained from the non-recursive non-systematic (conventional) convolutional encoder by feeding back one of its encoded outputs to its input.
Interleaver
For turbo codes, an interleaver is used between the two component encoders. The interleaver is used to provide randomness to the input sequences. The interleaver has the task to de-correlate the input sequence between the two RSCs. It generates out of Sys input bits a pseudo random re-ordered sequence of Sys bits. The de-correlation function of the interleaver is very important as only when the two input sequences Sys and Sys are most de-correlated, a powerful decoder can be realized. Note: In order to generate higher code rates, e.g. 1/2, puncturing can be used. Puncturing deletes bits according to some rule or algorithm and therefore reduces the number of output bits. The percentage of code-word to redundancy information gets reduced. However in turbo coding only the parity bits are punctured, never the systematic bits. Example: In order to generate from a 1/3 turbo coder a 1/2 bit sequence, the following scheme could be used: all systematic bits are transmitted, but the even numbered parity bits of P_1 and the odd numbered bits of P_2 are removed from the output stream N 1/(1+0.5 + 0.5) = 1/2.
3-4
MOTOROLA LTD.2002
Turbo Coding
Version 1 Rev 0
Turbo Coding
MOTOROLA LTD.2002
3-5
Version 1 Rev 0
Turbo Coder Structure

The scheme of a Turbo coder is a Parallel Concatenated Convolutional Code (PCCC) with two 8-state constituent encoders and one turbo code internal interleaver. Coding rate of a turbo coder is =1/3 (if the tail bits are not included) Turbo coding is block encoding suitable and efficient for code blocks. If turbo coding is selected and the number of input bits in the code block is less than 40, then filler bits are added to the beginning of the code block. The filler bits are transmitted and they are always zero. Note: After CRC attachment of the transport block, the block is called a coded block. After turbo coding the block is called an encoded block. The maximum code block size K for turbo coding is: 40 4 K 4 5114 If the number of coded bits exceeds 5114, code block segmentation is performed. The code blocks after segmentation are of the same size.
Trellis termination for Turbo coder

Trellis termination is done in order to reset the turbo coder after the coded block has been entirely encoded. Due to the recursive function in the RSC, an endless output bit sequence is generated even when the input sequence has ended. Therefore it is not possible to simply terminate the turbo coder by clocking in zero bits as it is done for convolutional coders. Trellis termination is performed by proper selection of tail bits. Because the states of the two RSC encoders will be usually be different after the data has been encoded, the tails for each encoder must be separately calculated and transmitted. The tail bits are generated for each encoder by throwing the two switches into the down position, thus causing the inputs to the two encoders to be indicated by the dotted lines. The tail bits are then transmitted at the end of the encoded frame according to x(k+1), z(k+1); x(k+2), z(k+2); x(k+3), z(k+3); x(k+1), z(k+1), x(k+2), z(k+2), x(k+3), z(k+3) The first three tail bits are used to terminate the first constituent encoder (upper switch of figure in lower position) while the second constituent encoder is disabled. The last three tail bits are used to terminate the second constituent encoder (lower switch of figure in lower position) while the first constituent encoder is disabled. As each RSC is a form of 1/2 convolutional coder, 2 times 6 output bits are generated by tail bits. When tail bits are taken into account the number of encoded bits is: 3K + 12 code rate = K / (3K + 12
Turbo Code internal Interleaver

The interleaver is a matrix with 5, 10 or 20 rows and between 8 and 256 column, depending on the size of the input word K. Data is read into the interleaver in a row-wise fashion. Intra-row permutations are performed on each row of the matrix. Next, inter-row permutations are performed. After that the data is read out from the interleaver in a column-wise fashion.
3-6
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-7
Version 1 Rev 0
Iterative Decoding Principle

The high error correction capability of turbo codes is mainly achieved through the special decoding method. A Turbo decoder consists in principle of two decoders. Both decoders exchange bit reliability information (so called soft decision values) about their decoded bits. The figure shows the principle of the iterative decoding. The input bit sequence, consisting of probabilities (soft values) for systematic bits and the corresponding parity 1 bits is firstly decoded in decoder 1. Decoder 1 imparts the soft decision values, which represent a measure of the probability that the input bit is a logical 0 or a logical 1, after they have been interleaved to decoder 2. Decoder 2, taking the soft decision values from decoder 1 into account, in addition to the soft values from systematic bits and the corresponding parity 2 bits, performs a much better error correction of the bit sequence and also generates soft decision values. These soft decision values are deinterleaved before they are imparted to decoder 1. This process can continue for several cycles (typically 3 to 6). In the decoder architecture, there are 2 decoders corresponding to two RSCs. The input to the first decoder are the corrupted systematic bits, the corrupted parity bit stream P_1 and the de-interleaved extrinsic information from the 2nd RSC. The inputs to the 2nd RSC are the corrupted interleaved systematic bit stream, the corrupted parity bit stream P_2 and the interleaved extrinsic information from the first decoder. The heart of the iterative decoding procedure in each component decoder is an algorithm that computes the Posteriori Probability of the information symbols or more generally a reliability value for each information symbol. The sequence of reliability values generated by a decoder is passed to the other one. In this way, each decoder takes advantage of the suggestions of the other one. To improve the correctness of its decision, each decoder has to be fed with information that does not originate from itself. The concept of extrinsic information was introduced to identify the component of the general reliability value, which depends on redundant information introduced by the considered constituent code. A equalizer (not shown in the picture) delivers soft information on the received data stream, i.e., it delivers probabilities of the received bits. These probabilities can be interpreted as containing the received bit value and the confidence value, which indicates how likely it is that this bit is correct. Decoder-1 then evaluates these probabilities and combines them with the parity-1 probabilities. This refines the soft information so the confidence of individual bit correctness is maximized. The refined probabilities are fed into decoder-2 with the information bits and the parity-2 bits, again producing enhanced soft information. After a predefined number of iterations (e.g. 3 to 6), the decoding process is completed, and the soft decision values are available at the output. When data is exchanged between the two decoders, the soft values are reordered to match the interleaving structure. This reordering is done with the interleaver and deinterleaver between the decoders. When the second decoder has finished, the next iteration is started. Again decoder-1 is activated using the soft information from the previous decoding as well as the information bits and parity-1 bits, and the second decoder is activated. Note: Prior to the first iteration the feedback information from decoder-2 is initialized to all zeros since decoder-2 has not yet acted on the data. [Trellis and Turbo Coding, Christian B. Schlegel; Lance C. Perez]
3-8
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-9
Version 1 Rev 0
TFRI - Transport Block Size Mapping

MAC-hs in the UE needs to be informed about the HS-DSCH transport block size. This is vital for the correct decoding of HS-DSCH. For HS-DSCH the transport block size is derived from the TFRI value signaled on the HS-SCCH. The mapping between the TFRI value and the transport block size for each modulation type is shown here and on the next page.
QPSK
For all transmissions of a transport block, the transport block size is derived from the TFRI value as specified in the figure. However there is an exception. In the case of retransmission, where the Node-B selects a combination for which no mapping exists between the original transport block size, the selected combination of channelization code set and the modulation type, then the transport block size index value signaled to the UE shall be set to 111111, i.e. k(i) = 63. The parameter k(i) corresponds to TFRI signaled on the HS-SCCH to the UE. The parameter k(0,i) is obtained from the table corresponding to the modulation type and the number of codes which are also signaled on the HS-SCCH to the UE. Let k(t) be the sum of the two values: k(t) = k(i) + (k0,i) The transport block size L[k(t)] can be calculated in the following way: IF k(t) < 40: L[k(t)] = 125 + 12 x k(t) ELSE L[k(t) = 296 x [(2085 / 2048) ^ k(t)]
3-10
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-11
Version 1 Rev 0

16-QAM
In the above example, the parameter TFRI which is signaled in part 2 of HS-SCCH is set to 29. The number of channelization codes is chosen to be 5. The number of channelization codes points to the parameter k(0,i) = 131 Accordingly k(t) can be calculated as: k(t) = 29 + 131 = 160 Accordingly L[k(t)] can be calculated as: L[160] = 296 x [(2085/2048) ^ 160] = 5193.6 = 5193 The formula for L[k(t)] allows the UE to calculate the transport block size efficiently, taking the number of channelization codes and the modulation type into account. The special TFRI = 63 is used for e.g. Full IR retransmission when no certain transport block size is applicable. At maximum 254 different TB_Indexs are possible to chose for the NodeB 192 + 62 Note: Some of the TB sizes are overlapping between QPSK and 16-QAM. Note: There is only one HS-DSCH transport block per TTI.
3-12
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-13
Version 1 Rev 0
HARQ Retransmission Comparison with Legacy Releases

The graphic opposite shows the important difference between Rel. 99 and HSDPA retransmission of RLC PDUs. The upper half of the picture shows the situation for the legacy releases while the lower half shows it for HSDPA in Rel. 5.
DCH
All Rel. 99 transport channels terminate at the RNC. Hence, the retransmission procedure for packet data is located in the SRNC, which also handles the connection for the particular core network. In RLC-AM the SRNC starts for every sent RLC-PDU a RLC retransmission polling timer. If the RLC status PDU positively or negatively acknowledging the RLC PDU in time a retransmission is sent. How many RLC retransmissions are performed for the same PDU depend on the bearer configuration, e.g. MaxDat parameter could be used to limit the RLC retransmission.
HS-DSCH
With the introduction of HSDPA additional intelligence in form of MAC-hs is installed in the NodeB. This way, retransmission can be controlled directly by the NodeB, leading to faster retransmission and thus shorter delay with packet data operation when retransmission is needed. The lower figure presents the difference. The RLC-PDU is sent via MAC-hs from NodeB towards UE. MAC-hs in NodeB monitors the successful transmission based on the fixed uplink timing of HS-DPCCH. If either the HS-DPCCH decoding fails or a Nack is received, then MAC-hs in NodeB schedules retransmission autonomously. In RLC-AM mode however the RLC of SRNC starts like in legacy UMTS releases a retransmission polling timer. If the RLC status PDU is not received in time to ack the RLC-PDU, the timeout of the RLC polling timer triggers the retransmission of the RLC-PDU. For MAC-hs it looks like an entire new transmission. With HSDPA the Iub interface between NodeB and SRNC (MAC-d) requires a flow control mechanism to ensure that NodeB buffers are used properly and that there is no data loss of RLC-PDUs due to NodeB buffer overflow. The MAC-hs in NodeB is responsible for scheduling retransmission. The RLC-PDU sent by SRNC to NodeB is embedded by MAC-hs into a so called MAC-hs-PDU. However, the SRNC still retains the Rel. 99 /Rel. 4 functionalities of RLC, such as taking care of retransmission (only in case of RLC-AM) in case of the HS-DSCH transmission from NodeB to UE fails after exceeding the maximum number of physical layer retransmission or a MAC-hs PDU release timer T1 expires. The key functionality of the MAC-hs is to handle the HARQ functionality and scheduling as well as priority handling. Ciphering is done in any case in the RLC layer to ensure that the ciphering mask stays identical for each retransmission. This enables physical layer combining of retransmission. Note: The type of scheduling to be carried out in NodeB is not defined in 3GPP standardization. Only some parameters, such as discard timer or scheduling priority indication can be used by RNC to control the handling of an individual user. In chapter 4 we explain these parameters and their functions in full detail.
3-14
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-15
Version 1 Rev 0
Redundancy Version and Constellation Version

X(RV) is signaled to the UE via HS-SCCH immediately before the HS-DSCH data block is transmitted. This parameter comprises the redundancy version and, in the case of 16-QAM, the constellation version. The redundancy version contains the self-decodable indication s and redundancy parameter r. NOTE Note: A transmission or retransmission is self-decodable, when the systematic bits of the encoded blocks are transmitted. The systematic bits are those bits which are necessary to decode the data block. This means the data block must at least comprise the systematic bits. The remaining bits, if present, represent the added redundancy
Chase combining: Every transmission and retransmission of HS-DSCH data blocks must be always self-decodable. Therefore s = 1 and r = 0 for all HS-DSCH data blocks. As there is only one value allowed for r, chase combining contains always the same redundant parity bits. Partial IR: Also when partial IR is performed, every transmission and retransmission must be always self-decodable. Therefore s = 1 but different redundant bits of the parity information are added. The different parity bits are selected by the parameter r. Full IR: Compared to partial IR, full IR is only used for retransmission. The initial transmission must be self-decodable in order to transmit the systematic bits. Bit Rearrangement only for 16-QAM: The positions I2 and q2 are less reliable therefore in retransmission the bits are rearranged in order to have equal error probability among transmission and retransmission.
3-16
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-17
Version 1 Rev 0
HSDPA Category and Reference Combinations

With the introduction of the HSDPA concept into the Rel. 5 specifications, a new generation of UE capability classes has been introduced.
HS-DSCH Category
New capability parameter introduced in Rel. 5. Category 11 and 12 only support QPSK, all other categories support both QPSK and 16-QAM.
Maximum number of HS-DSCH multi codes

This parameter indicates the maximum number of channelization codes with SF=16 that can be simultaneously received by the UE. At least five multi codes must be supported in order to facilitate efficient multi-code operation.
Minimum inter-TTI interval

This parameter defines the distance in time from the beginning of a TTI to the beginning of the next TTI that can be assigned to the same UE. E.g. if the allowed interval is 2 ms, this means that the UE can receive HS-DSCH data every 2 ms.
Maximum Number of HS-DSCH Transport Channel Bits received within HS-DSCH TTI
This parameter defines the maximum transport channel bits which can be received by the UE within 2 ms interval.
Total Number of Soft Channel Bits over all HARQ Processes

This parameter defines the max number of soft channel bits over all HARQ processes. When explicit signaling is used, C-RNC configures process memory size for each HARQ process so that the following criterion must be fulfilled in the configuration: Total number of soft channel bits for HS-DSCH 7 sum of process memory size of all the HARQ processes
Maximum Number of RLC-AM Entities

This parameter defines the maximum number of RLC-AM entities supported by UE.
Total RLC-AM and MAC-hs Buffer Size [KBytes]

This parameter informs the network about the layer 2 buffer capability in UE. This buffer size can be dynamically shared by RLC-AM and MAC-hs entities. Note that a low-end HSDPA UE will support a maximum of 5 simultaneous HS-DSCH codes, and the minimum distance between the starting points of two successive data packets is 3 TTIs (i.e. 6 ms). Such a UE will support a max. of 7300 bits in each TTI and thus belong to the 1.2Mbit/s class. The number of soft channel bits will impact the UE receiver performance when HARQ is employed. A UE with low number of soft channel bits will not be able to support IR for the highest peak rates and its performance will thus be slightly lower than for a UE supporting a larger number of soft channel bits.
3-18
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-19
Version 1 Rev 0
Number of HARQ Processes versus IR Memory

Here we would like to give some considerations that the maximum data rate may only be achieved with Chase Combining. The issue is that the total number of soft channel bits over all HARQ processes must be split up over the number of HARQ processes. The HARQ processes perform soft combining and may need to cater for retransmission with Full and / or Partial IR. Whenever retransmission with IR is performed, then UEs IR soft-buffer must not be overloaded. Therefore the NodeB has to consider the number of HARQ processes in conjunction with the total size of soft-buffer per UE. We chose for our example again UE with category 6: 67200 bits total soft buffer capacity Number of HARQ processes = 4 Total transmission capacity with 5 HS-PDSCHs: QPSK: 2 bits/symbol x 240 ksps x 5 x 2 ms = 4800 bits 16-QAM: 4 bits/symbol x 240 ksps x 5 x 2 ms = 9600 bits Maximum TB size = 7298 bits Total number of coded bits: (7298 + 24 CRC) x 3 + 12 = 21978 bits 14652 Parity P1 and P2 bits Soft-buffer capacity per HARQ process: 67200 / 4 = 16800 bits Note for a consecutive subframe allocation, taking also retransmission into account, more than 4 HARQ processes are necessary.
Chase Combining
QPSK: no problem the total transmission capacity is only 4800 bits with 5 HS-PDSCHs so the UEs soft-buffer cannot be overloaded 16-QAM: no problem the total transmission capacity with 9600 bits is still below the maximum HARQ process memory 16-QAM code rate R: 7298 / 9600 bits = 0.76 ~
IR Retransmission
QPSK: no problem the total transmission capacity is only 4800 bits with 5 HS-PDSCHs so the UEs soft-buffer cannot be overloaded 16-QAM: issue the total number of bits of 21978 exceeds UEs soft-buffer capacity puncturing of parity P1 and P2 bits down to 9474 bits Alternative for 16-QAM: (16800 bits 12) / 3 24 = 5572 bits With TB size 8 5572 bits UEs total soft-buffer capacity gets never exceeded when Partial IR or Full IR are used for retransmission. If the amount of soft-buffer is not an issue, then IR is superior compared to chase combining. However, e.g. for UE category 10, the available soft-buffer capacity with 6 HARQ processes is only 28800. If the biggest TB size of 27952 is chosen, then only chase combining makes sense as no further parity bits can be retransmitted when UEs soft-buffer is only marginal bigger than UEs max. physical transmit capacity.
3-20
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-21
Version 1 Rev 0
HARQ Information
HARQ Information
The UE configures the MAC-hs entity with the number of HARQ processes indicated in IE "Number of Processes". The UE assigns to each of these HARQ processes IDs going from 0 to "Number of Processes" 1 HARQ Memory Partitioning Information: The first instance of the parameter corresponds to HARQ process with identifier 0, the second instance to HARQ process with identifier 1, and so on.
Memory Partitioning is set to Implicit

UE partitions the soft memory buffer in the MAC-hs entity equally among the processes configured by the IE Number of Processes. Note: In the event that the division of the soft memory buffer results in a non-Integer value the partition memory size is rounded down to the nearest Integer value.
Memory Partitioning is set to Explicit

The UE partitions the soft memory buffer in the MAC-hs entity according to the IE "Process memory size" assuming that the order in the list follows the order in the HARQ process IDs. The first instance of the parameter corresponds to HARQ process with identifier 0, the second instance to HARQ process with identifier 1, and so on Note: If the UE capability "Total number of soft channel bits in HS-DSCH" is exceeded with this configuration, the UE should report INVALID_CONFIGURATION. The UE capability is specified by its category ranging from 14400 (category 11, QPSK) to 172800 (category 10, 16-QAM). Note: In chapter 4 we explain in detail the functionality of HARQ process inside MAC-hs.
3-22
MOTOROLA LTD.2002
HARQ Information
Version 1 Rev 0
HARQ Information
MOTOROLA LTD.2002
3-23
Version 1 Rev 0
HARQ Principle in HSDPA

NodeB Side
There is one MAC-hs entity in the UTRAN for each cell that supports HS-DSCH transmission. The MAC-hs is responsible for handling the data transmitted on the HS-DSCH. Furthermore it is responsible for the management of the physical resources allocated to HSDPA. MAC-hs receives configuration parameters from the RRC layer. Priority Queue: MAC-hs SDUs come in from different priority queues. Based on status reports from uplink Ack/Nack signaling either new transmission or retransmission is determined for a individual UE. Further the queue ID and a TSN (transmitted sequence number) for each new MAC-hs PDU being serviced is determined. Retransmitted MAC-hs PDUs keep their TSN. N-Channel Distribution: New MAC-hs SDUs are routed to the proper TX Buffer of the selected HARQ process. Based on the status reports from HARQ processes the scheduler determines if either a new transmission or a retransmission should be made. A new transmission can however be initiated on a HARQ process at any time. Based on a delay attribute provided by upper layers, the scheduler may decide to discard any out-of-date MAC-hs SDU. HARQ Entity handles the hybrid ARQ functionality for one user. One HARQ entity is capable of supporting multiple instances (up to 8 HARQ processes N= 8) of SAW (stop and wait) HARQ protocols. There is one HARQ process per HS-DSCH per TTI per UE. For every HARQ process there is a TX buffer which allows for immediate retransmission with or without different redundancy version once the Ack/Nack has been received for the corresponding HS-DSCH data transmission. As indicated by the picture, while waiting for the Ack/Nack of a certain MAC-hs PDU belonging to a certain HARQ process, another MAC-hs PDU can be transmitted to the same UE. Therefore for each HARQ process per UE a TX buffer is required in NodeB.
UE Side
HARQ entity is responsible for generating Ack/Nacks of the data in the corresponding HARQ process and schedule the positive or negative acknowledgement for transmission relative to the reception of data ( HARQ control N). The detailed configuration of HARQ protocol is provided by RRC. In order to combine multiple retransmissions of the same MAC-hs PDU belonging to a certain HARQ process, the UE need to have a buffer per HARQ process. Note: After channel decoding, the CRC check determines if the HS-DSCH transport block is correct or erroneous. Reordering Queue distribution function routes the MAC-hs PDUs to the correct reordering buffer based on the queue ID. Reordering entity & Disassemble: The reordering is done according to the received TSN in the MAC-hs PDU. MAC-hs PDUs with consecutive TSNs are delivered to the disassembly function upon reception. MAC-hs PDUs are not delivered to the disassembly function if MAC-hs PDUs with lower TSN are missing. There is one reordering entity for each queue ID configured at the UE.
3-24
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-25
Version 1 Rev 0
HS-DSCH HARQ Functionality

Turbo Coding
The physical-layer Hybrid-ARQ functionality is an extension of the Rel. 99 rate matching. The Hybrid-ARQ functionality matches the number of bits at the output of the turbo coder to the total number of bits of the HS-DSCH physical channels. The turbo coder increases the transport block data rate by factor 3. The bit stream N(TTI) contains the encoded bits including the ones from Trellis: x(1), z(1), z(1), x(2), z(2), z(2), , x(k), z(k), z(k), x(k+1), z(K+1), x(K+2), z(K+2), x(K+3), z(K+3), x(K+1), z(K+1), x(K+2), z(K+2), x(K+3), z(K+3) which are split up into 3 sequences each comprising of N(TTI)/3 bits.
Bit Separation
This unit separates the bits stream in systematic bits (x(1), (x(2),..., x(k), x(k+1), z(k+2) x(k+1), z(k+2)), parity 1 bits (z(1), z(2), ..., z(k), z(k+1), x(k+3) z(k+1), x(k+3)) and parity 2 bits (z(1), z(2), ..., z(k), x(k+2), z(k+3) x(k+2), z(k+3)).
1st RM
The first rate-matching stage is identical to the Rel.99 rate-matching functionality except that the number of output bits does not match to the number of physical-channel bits available in the HS-DSCH TTI. Instead the first rate matching block is used to adjust the number of available encoded bits N(TTI) to the virtual UEs buffer size. Note: The first rate matching block is always transparent for the systematic bits therefore only parity-1 and parity-2 bits are punctured. If N(TTI) UE soft-buffer capability (N(IR) > 3N), the first rate matching stage is also transparent for parity-1 and parity-2 bit streams.
2nd RM
The second rate matching stage matches the number of bits after first rate matching to the number of physical channel bits available in the HS-DSCH TTI. The second rate matching also uses the Rel. 99 rate matching algorithm. However, the rate matching only considers bits that have not been punctured by the first rate matching stage and the rate matching parameters used in a particular transmission are controlled by the RV parameter.
RV
The HARQ functionality is controlled by the parameter RV, i.e. the exact set of bits at the output of the physical-layer Hybrid-ARQ functionality depends on the number of input bits, the number of available physical channel bits (number of HS-PDSCHs and modulation type), and the RV parameter. Note: The number of RVs allowed for HS-DSCH is limited to 8. Parameter RV decides about the different redundancy versions for IR.
Bit Collection
The HARQ bit collection is achieved using a rectangular interleaver. The size of the interleaver is determined by the modulation type and the number of encoded and rate matched bits. The purpose of the bit collector is to transmit the systematic bits on the more reliable positions of a 16-QAM symbol. A perfect mapping exists only for code rate 1/2. Otherwise it is just best effort.
3-26
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-27
Version 1 Rev 0
HARQ Parameters for Retransmission

The picture is shows which parameters remain the same for the retransmission and which parameter are changeable HARQ.
Transmission Parameters = Retransmission Parameters

The TBS (transport block size) is the same as for the initial transmission. This means that, for a retransmission, the number of transport channel bits to be transmitted is the same as for the initial transmission. N(TTI) remains the same.
Transmission Parameters <> Retransmission Parameters

Modulation type may change between QPSK and 16-QAM depending on the UEs CQI reporting and/or DL DPCH power control status. In the long run also the Ack/Nack ratio may influence the selection of the modulation type. The channelization-code set, including the number of HS-PDSCHs available for HS-DSCH transmission (channelization code size) is subject to changes. Downlink Transmission power may vary between the HS-DSCH TTIs depending UEs CQI reporting and/or DL DPCH power control status. The N(CH) bits (number of channel bits after rate matching) may differ compared to the initial transmission even if the sum of N(CH) is the same. N(CH) (set of channel bits consisting of systematic, parity 1 and parity 2 bits) differs only for retransmission in case of IR only. In case of 16-QAM, the mapping of the output bit sequence is rearranged to provide for equal error probability among transmission and retransmission. Constellation rearrangement averages the bit reliability in retransmission by changing the mapping of four consecutive bits onto symbols. Constellation rearrangement gain is only achieved after retransmission and provides increasing advantages with increasing similarity of the transmitted bits especially for HARQ type III (e.g. chase combining). Thus the gain for full IR with a high initial code rate is not so strong. In case of IR, partial IR or full IR, different versions of parity bits are used in retransmission thus reducing the initial code rate with every retransmission.
3-28
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-29
Version 1 Rev 0
Rate Matching Tasks
Rate Matching Tasks

First Rate Matching
The first RM stage matches the number of encoded bits to the UEs soft-buffer capability. The available UEs virtual buffer size is signaled via CQI reporting to the NodeB while running HS-DSCH transmission. In addition the NodeB is informed at the beginning of HSDPA transmission via e.g. Radio Link Setup and the process memory size IE about UEs soft-buffer capability. The HARQ memory partitioning IE provides the total number of soft channel bits which are either equally partitioned between all HARQ processes according to RRC or HARQ memory partitioning is explicit configured by RRC. Then the first instance of the parameter corresponds to HARQ process with identifier 0, the second instance to HARQ process with identifier 1, and so on.
Second Rate Matching

N(CH) = N(data) The second RM stage ensures that the encoded bits N(CH) match the available physical channel capacity indicated by N(data). N(data) is the number of bits available to the HS-DSCH in one TTI. N(data) depends on the modulation type which delivers N(data1) bits per TTI and the number of HS-PDSCHs used for multi code operation (available channelization code size P which ranges from 1 to 15): N(data) = P x N(data1) P represents the number of HS-PDSCHs N(data1) for 16-QAM: 960 kbit/s x 2 ms = 1920 bits N(data1) for QPSK: 480 kbit/s x 2 ms = 960 bits Bit Rearrangement and Redundancy Version Selection As already mentioned, the modulated symbols in 16-QAM contain 2 highly reliable positions namely i1 and q1 and two less reliable positions i2 and q2. In order to average the bit reliability after retransmission, the positions of i1,q1,i2 and q2 are modified according to bit rearrangement parameter b. The redundancy version selection is only changeable for IR (r = 1..r(max)). Chase combining always uses only one redundancy version (s=1 and r=0).
3-30
MOTOROLA LTD.2002
Rate Matching Tasks
Version 1 Rev 0
Rate Matching Tasks
MOTOROLA LTD.2002
3-31
Version 1 Rev 0
First RM Stage
First RM Stage
HARQ first stage rate matching for the HS-DSCH transport channel is done with the general method as described in the picture with the following specific parameters: The maximum number of soft channel bits available in the virtual IR buffer is N(IR) which is signaled via RRC for each HARQ process. The number of coded bits in a TTI before rate matching is N(TTI) this is deduced from information signaled via RRC as well and parameters signaled on the HS-SCCH for each TTI HARQ processing and physical layer storage occurs independently for each HARQ process currently active. If N(IR) is greater than or equal to N(TTI) (i.e. all coded bits of the corresponding TTI can be stored) the first rate matching stage is transparent. Note that no repetition is performed in 1st RM stage. The 2nd RM stage matches the encoded bits of the TB to the available channel capacity N(data). This function is explained later on. If N(IR) is smaller than N(TTI) the parity bit streams are punctured by setting the rate matching parameter 3N(TTI) = N(IR) N(TTI). NOTE The negative value is expected when the rate matching implements puncturing. Bits selected for puncturing which appear as x(m) in the algorithm depicted above shall be discarded and not counted in the totals for the streams through the virtual IR buffer.
Puncturing Algorithm
The variables e(ini), X(i), e(plus) and e(minus) must be initialized accordingly. X(i) represents the number of parity bits before 1st RM. As X(i) = N(TTI)/3 and N(TTI) is always a multiple of 3 the result is always an integer. e(ini) is the start value in the beginning of the algorithm which sets the variable e =X(i) at the beginning only. e(minus) decrements e in order to get e 8 0. Only when e 8 0 the corresponding bit can be deleted. x(m) represents the bit position of the bit being deleted once e 8 0. The parameter m starts with first bit position and is incremented after every loop cycle. e(plus) increments e in order to push e above zero again after a bit has been determined by the loop for puncturing 3N(TTI) determines the number of bits to be punctured in total. As parity-1 and parity-2 have the same amount of bits, 3N(TTI)/2 bits must be deleted for each parity stream. At the end of the puncturing process, the algorithm delivers the amount of N(TTI)/3 - 3N(TTI)/2 bits for parity-1 and parity-2
3-32
MOTOROLA LTD.2002
First RM Stage
Version 1 Rev 0
First RM Stage
MOTOROLA LTD.2002
3-33
Version 1 Rev 0
Second Rate Matching Function

The figure shows how the variables for the second RM function are calculated. HARQ second stage RM for the HS-DSCH transport channel is done with the general method also used for legacy release channels with the following specific parameters: The parameters of the second RM stage depend on the value of the RV parameters s and r. The parameter s can take the value 0 or 1 to distinguish self-decodable (s = 1) and non self-decodable (s = 0) transmissions. The parameter r with range 0 to r(max) changes the initial error variable e(ini) in the case of puncturing. In case of repetition both parameters r and s change the initial error variable e(ini). The parameters X(i), e(plus) and e(minus) are calculated as shown in the figure. The goal of the e(ini) variation parameter is to enable robust retransmission by orthogonalizing available redundancy versions that contain the same self-decodability parameter s. The initial e(ini), increment e(plus) and decrement e(minus) value of variable e in the rate matching pattern determination algorithm (shown further below) are based on the RV parameters s and r. The parameter s can take the value 0 or 1 to distinguish between prioritizing systematic and parity bits. The parameter r changes the initial error variable e(ini) in the case of puncturing. In the case of repetition, both parameters r and s change the initial error variable e(ini). Denote the number of bits before second RM as N(sys) for the systematic bits, N(p1) for the parity-1 bits and N(p2) for the parity-2 bits, respectively. Denote the number of physical channels used for the HS-DSCH CCTrCH by N(data). For N(data) 4 N(sys) + N(p1) + N(p2) puncturing is performed in the second rate matching stage. The number of transmitted systematic bits in a retransmission is N(t,sys) = min {N(sys), N(data)} for a transmission of self-decodable type and N(t,sys) = max {N(data)-(N(p1) + N(p2))} in the non self-decodable case. For N(data) > N(sys) + N(p1) + N(p2) repetition is performed in the second rate matching stage. The RM parameter e(ini) is calculated for each bit stream according to the RV parameters r and s using the equations shown in the figure. Note: For the modulo operation the following clarification is used: The value of (x mod y) is strictly in the range of 0 to y-1 (i.e. -1 mod 10 = 9).
3-34
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-35
Version 1 Rev 0
RM Pattern Determination
The figure shows the algorithm determining which bit positions are to be punctured or repeated. The variables X(i), e(ini), e(plus) and e(minus) are calculated earlier. X(i) is the number of bits in each sequence {systematic bits}, {parity-1 bits} and {parity-2 bits} m is the loop variable which runs from 1 to X(i) and determines the bit number which is to be deleted or punctured.
Bit Repetition
A similar repetition rate in all bit streams is achieved by setting the number of transmitted systematic bits to: N(t,sys) = N(sys) x N(data)/[N(sys) + 2 x N(p1)] The number of parity bits in a transmission is: N(t,p1) = [N(data) N(t,sys)] / 2 N(t,p2) = [N(data) N(t,sys)] / 2 Example: N(sys) = 30; N(p1) = N(p2) = 30; N(data) = 100 N(t,sys) = 30 x 100/[30 + 2 x 30] = 30 x 100 / 90 = 30 x 10 / 9 = 34 N(t,p1) = [100 34] / 2 = 66./2 = 33 N(t,p2) = [100 34] / 2 = 66./2 = 33 So 4 Bits of the systematic bits are repeated as well 3 bits each of both parity bits are repeated.
3-36
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-37
Version 1 Rev 0
Example for Initial Transmission Self-decodable with 1st RM

Here we would like to give an idea how the initial transmission and subsequent retransmission could look like for full IR and chase combining. On top of this we show transmission and retransmission for two different transport block set sizes. UE Capability: e.g. category; 4 HARQ Processes total number of soft memory = 38400 / 4 = 9600 bits IR buffer per HARQ process 5 HS-PDSCHs; QPSK 4800 bits / 2 ms; X(RV) = 0 a) CQI = 11 TBS = 11 MAC-hs decides for TFRI(a) = 1537 b) CQI = 15 TBS = 3319 MAC-hs decides for TFRI(b) = 3379
Example for Initial Transmission-Self-decodable with 1st RM

Due to turbo encoding the transport block size triples. For the exact calculation of coded bits, 24 CRC bits must be added and 12 bits after turbo encoding must be added because of trellis termination. Note that N(TTI) is always a multiple of 3. Due to turbo encoding the transport block size triples. For the exact calculation of coded bits, 24 CRC bits must be added and 12 bits after turbo encoding must be added because of trellis termination. Note that N(TTI) is always a multiple of 3. The turbo encoded bits are then separated into systematic bits N(TTI) / 3, parity 1 N(TTI) / 3 and parity 2 bits N(TTI) / 3. The correct order of bits fed into the 1st RM stage is maintained by the bit separation function. As already mentioned, the 1st RM stage is always transparent for the systematic bits and only applies to parity-1 and parity-2 bits if the sum of 9(N(sys),N(p1)N(p2)) exceeds UEs virtual IR buffer capacity. The NodeB is informed by C-RNC about UEs physical layer category allowing MAC-hs to estimate the number of soft buffer capability per HARQ process. Beside this the UE always signals via CQI the available N(IR). In our example we can see that for case N(TTI) > N(IR) and therefore 311 parity-1 and 310 parity-2 bits need to be punctured. Calculation: 9600 (3407 + 2 x 3407) / 2 = - 621 RM_P1_1 and RM_P2_1 need to puncture 311 bits respectively 310 bits in order to match the encoded Bits to N(IR) of 9600 bits per TTI of 2 ms. The 2nd RM stage matches N(CH) bits of the encoded TB to the physical channel capacity N(data). Some more bits need to be punctured from parity-1 and parity-2 bits as N(CH) > N(data). Note: In case of initial transmission of a TB, the entire amount of systematic bits need to be transferred s = 1 self-decodable transmission. Therefore it can be seen that only parity-1 and parity-2 bits are subject to puncturing again. In case a) all bits of the turbo encoded TB can be sent. Therefore the code rate R = 1/3. If retransmission is requested, the same encoded bits can be sent again thus the receiver can perform chase combining. The systematic bits comprise of 71% of N(data), so only 29% of the parity bits can be sent with the initial transmission. If radio conditions are favorable, the remaining 4800 parity bits can be discarded from the transmitter queue and the TB was sent with a code rate R = 3379 / 4800 = 0.7. However, if there is retransmission necessary, the remaining parity bits will have to be sent. Calculation: N(data) = 100% systematic bits in % = 3407/4800 = 71%
3-38
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-39
Version 1 Rev 0
Example for Retransmission Non-self-decodable

The depicted example shows a typical case for full IR. In the original transmission all systematic bits were sent. After a Nack from the UE, only the remaining parity bits will retransmitted.
Full IR
The parameter s = 0 indicates non-self-decodable transmission and parameter r = 1 indicates which parity bits will be transmitted according to the 2nd RM algorithm. In the original transmission 22% of the parity bits were already transmitted. Calculation: (696 + 697) / (3097 + 3096) = 22% In the retransmission the remaining 78% parity bits can be transmitted. Calculation: 4800 / (3097 + 3096) = 78% The significant sign of full IR is that the systematic bits are not retransmitted as shown in the picture. Below we would like to explain shortly how the retransmission would like if partial IR is used.
Partial IR
In this case also the retransmission contains all of the systematic bits thus always only 22% of the parity bits can be transmitted. Calculation: (696 + 697) / (3097 + 3096) = 22% Thus it takes much longer until all parity bits are transmitted. Theoretically it would take 5 transmissions in order to transmit all of the parity bits. Each retransmission would contain 3407 systematic bits. The advantage of full IR becomes obvious now. Higher data rates get more effectively retransmitted however the disadvantage is that the systematic bits are only contained in the original transmission. So the UE must be able to successfully decode the TB out of the corrupted original transmission and the parity bits.
3-40
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-41
Version 1 Rev 0
Example for Chase Combining

The figure shows the case for original transmission and retransmission using chase combining. All bits of the turbo encoded TB (systematic and parity bits) can be sent in one shot. Therefore the code rate R = 1/3. If retransmission is requested, the same encoded bits are sent again thus the receiver performs chase combining and the code rate remains the same. The parameter s is set in every transmission (including retransmission) to 1 every transmission is self-decodable The parameter r is set in every transmission (including retransmission) to 0 which means only one redundancy version.
Overview over transmission capacity considerations

The systematic bits must always be included in the original (initial) transmission) of every new HS-DSCH TB. The data rate can reach at maximum the physical channel transmission capacity. This means at initial transmission only the systematic bits are transmitted. If retransmission are necessary, then 50% of the parity bits are sent full IR case. Is yet another retransmission required, then the remaining 50% of the parity bits are sent. Please bear in mind that N(CH) corresponds to 100% N(data) in all three transmissions. N(data) = 100% = N(t,sys); N(t,p1) = N(t,p2) = N(data) If e.g. the data rate consists of 60% of N(data) systematic bits occupy 60% of N(data), then 1/3 of the parity bits can already be sent within the initial transmission. 2 x parity bits = 120% of N(data) 1/3 = 40% If retransmission is necessary, then the remaining parity bits can be sent which only occupy 80% of N(data) provided that the physical channel capacity does not change. The unused space in the HS-DSCH TTI is filled by RM with repetitions 20% of the parity bits are repeated.
3-42
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-43
Version 1 Rev 0
2nd RM Parameter Calculation Self-Decodable

The figure shows a basic example for rate matching calculation. In order to keep the math simple, only a very small transport block size of 8 bits was chosen. Here we just want to demonstrate how the formulas apply to calculate the different variables for rate matching.
Turbo Coding with Transparent 1st RM stage

The turbo encoder gets a TB consisting of 8 bits. Now for simplicity reasons, the 24 bits CRC are skipped thus the output of the turbo coder is 3 x 8 bits = 24 bits. The bit separation function splits the bit stream into 8 systematic bits and 2 times 8 parity bits (parity-1 and parity-2 bits). These three input sequences are fed into the 1st RM stage. The systematic bits are never punctured by the 1st RM stage. As the virtual IR buffer indicated by UE via CQI reporting is big enough, none of the parity bits need to be punctured. The three bit sequences denoted as N(sys) = 8, N(p1) = 8 and N(p2) = 8 are now fed into the 2nd RM stage. The 2nd RM stage now has the task to reduce the amount of parity bits so that the bit stream of N(CH) matches N(data). The HS-PDSCH transmission capacity indicated with N(data) is set to 20 bits. Thus 4 parity bits have to be punctured. The reason why none of the systematic bits get punctured is due to the parameter s = 1which indicates self-decodable transmission.
X(i), e(plus) and e(minus) Calculation

The values for the variable X(i) are derived after the 1st RM stage. In our case the 1st RM stage is entirely transparent thus N(sys) = 8, N(p1) = 8 and N(p2) = 8. The values for the variable e(minus) have to be calculated in dependency of parameter s and N(data). As self-decodable transmission is chosen and therefore all systematic bits must be transmitted, e(minus) is set to 0. The values for the variable e(minus) of parity-1 and parity-2 bits heavily depend on the difference between N(data) and N(t,sys) which is nothing less the remaining bit space after N(data) has been filled up with all systematic bits. The difference between physical transmission capacity N(data) and N(t,sys) results in 6 bits for each parity stream. Thus 2 bits per parity stream needs to be punctured. In order not to puncture the same parity bits in parity-1 and parity-2 bit streams per redundancy version, the value of e(minus) for parity-1 is always doubled compared to parity-2. The variables e(minus) and e(plus) are used in the RM loop. The variable e(minus) is used to decrement bit position variable e and e(plus) is used to increment e again. As e(minus) is doubled for parity-1, also e(plus) needs to be doubled.
Initial Setting of e and Redundancy Version r

For demonstration purpose, we calculated the value of variable e(ini) for all four redundancy versions. Actually for our example it does not matter if QPSK or 16-QAM is used. The modulo operation is quite important especially to understand what happens at negative input!
3-44
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-45
Version 1 Rev 0
2nd RM Bit Position Calculation Self-Decodable

Here we now determine the bit positions going to be punctured. The red color coding indicates which bit-positions are punctured. As we decided to go for self-decodable transmission, s = 1. All values for the variables are copied from the previous page in order to ease the understanding. Also the RM loop is depicted in order to see how e is going to change. The variable m represents the bit positions and therefore counts from 1 to 8 ( TB = 8). The RM loop has the task to determine always which bit position e (up to m) has to be removed. The value for e(ini) is obtained from the previous page. It can be seen that e(ini) differs with the redundancy version parameter r. Note that e(ini) is always different between parity-1 and parity-2 in order to have different start values for the RM loop.
Systematic RM
As e(minus) is always 0, no bits are punctured e never becomes less or equal to 0. The variable e remains constant and positive; e = e(ini) for every bit position.
Parity_1 RM
Here the variable e gets decremented for every m by e(minus). We always indicate the result of the subtraction per m. Once e becomes less or equal 0, the very bit position indicated by m needs to be punctured. Thus we mark that row with red color. According to the RM loop algorithm, whenever e becomes less or equal to 0, e has to be incremented by e(plus) in order to continue.
Parity_2 RM
The bit position calculation for parity-2 bits is basically the same as for parity-1. Only the loop variables e(ini), e(plus) and e(minus) differ in order to get different results compared to parity_1 RM. Thus different bit positions are punctured. Note: For retransmission with s = 1 and always the same redundancy version, e.g. r = 0, this scheme represents HARQ III (chase combining).
3-46
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-47
Version 1 Rev 0
2nd RM Self-Decodable Transmission followed by Full IR (1)

The figures show a very aggressive second RM. Out of originally 24 bits (TB size = 8 bits) only 13 bits can be transmitted R = 8 / 13 = 0.62 The first transmission must be self-decodable. As can be seen by the RM algorithm, e(minus) variable is set to zeoro. This means non of the systematic bits will be punctured in the first transmission. The RM calculation for parity bits shows that only 2 bits of parity_1 stream and 3 bits of parity_2 stream are transmitted. On the next pages the possible retransmission is shown with Full IR. It must be pointed out that with Full IR retransmission method, the TB size does not make sense to be indicated in retransmission as only parity bits are transmitted. Therefore the TB size may be set to 63 in possible retransmission.
3-48
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-49
Version 1 Rev 0

The figures show a very aggressive second RM. Out of originally 24 bits (TB size = 8 bits) only 13 bits can be transmitted R = 8 / 13 = 0.62 The first transmission must be self-decodable. As can be seen by the RM algorithm, e(minus) variable is set to zero. This means none of the systematic bits will be punctured in the first transmission. The RM calculation for parity bits shows that only 2 bits of parity_1 stream and 3 bits of parity_2 stream are transmitted. On the next pages the possible retransmission is shown with Full IR. It must be pointed out that with Full IR retransmission method, the TB size does not make sense to be indicated in retransmission as only parity bits are transmitted. Therefore the TB size may be set to 63 in possible retransmission.
3-50
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-51
Version 1 Rev 0
2nd RM Parameter Calculation Non-Self-Decodable

The figure shows a basic example for full IR. The same number of bits for the TBS is used like in the previous example. Note: The systematic bits have been already transmitted in the original transmission. This is now a sub-sequent retransmission with parity bits only. As can be seen, N(data) is less than the sum of parity bits, therefore the 2nd RM stage need to remove 3 parity bits in total in order to match N(CH) to N(data). Which bits get removed per parity stream is shown on the next slide. As s = 0 (non-self-decodable), none of the systematic bits are transmitted. The values for the different variables are calculated according to the same algorithm like in the self-decodable example and thus not further elaborated here.
3-52
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-53
Version 1 Rev 0
2nd RM Bit Position Calculation Non-Self-Decodable

As the systematic bits are all removed in non-self-decodable retransmission, they are not shown here. Anyway the parameter e(minus) for the systematic RM is 8 thus variable e is always 0. This means that all systematic bits get removed as e(plus) is also only 8. The bit removal calculation follows the same algorithm like in the previous example. The only difference is that the variable e(minus) for Parity_2 RM has a non-integer value now. Again we calculated the punctured bit positions for all four redundancy versions. The transmitted bit positions per r are depicted on the right side of the picture. It can be seen which bit positions are removed for parity-1 and parity-2 bit stream.
3-54
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-55
Version 1 Rev 0
Rate Matching Exercise

Please calculate the bit positions being removed for a non-self-decodable retransmission?
3-56
MOTOROLA LTD.2002
Version 1 Rev 0
Rate Matching Exercise Exercise Solution
MOTOROLA LTD.2002
3-57
Version 1 Rev 0
Comparison between Full and Partial IR

The code rate R expresses the ratio of protection, and how much redundancy is added to the user bits respectively. A code rate close to 1 means that there is actually almost no redundant information added. Thus such aggressive transmissions are performed under good ratio conditions and high user data rate requests. Full and Partial IR allow to reduce the code rate compared to chase combining. However, chase combining requires minimum complexity and buffer size. Initial transmission and retransmission are identical when using chase combining method. Chase combining offers time diversity and soft combining gain taking the SNR of each (re-) transmission into account (i.e. the energy accumulation effect for each bit). The code rate r is defined as: R = (number of bits before channel coding) / (number of bits after channel coding & puncturing) Note: The number of bits before channel coding (turbo coding) matches the systematic bits (in our example: 3 bits (blue color)) As an example, the figure compares partial and full IR for code rates R = 3/4 and R = 1/2 in a pictorial way. The code rate is re-calculated after each retransmission. The first row in each block signifies the total number of bits after turbo coding with R = 1/3. Due to the turbo coder code rate of 1/3, twice as much parity bits (yellow) are available than systematic bits (blue). The following four rows show which bits are sent in each transmission N(t). For partial IR the systematic bits are always repeated, simply different parity bits are sent.
R = 3/4
For high initial code rates it is obvious that it takes partial IR several retransmissions until all parity bits have been sent once. R(eff) is slowly decreasing and only after six transmissions all bits have been sent (3/4, 3/5, 3/6, 3/7, 3/8, 3/9) Full IR allows decreasing the effective code rate more rapidly. As depicted in the figure, all bits have been sent once already after three transmissions.
R = 1/2
For R = 1/2, however, there is no longer a decoding gain advantage for full IR, since partial IR allows also to send all available bits within two transmissions. Detailed investigations have shown, that for HSDPA full IR is favorable for initial code rates R > 1/2, while partial IR should be used for lower initial code rates, e.g. R8 1/2.
3-58
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-59
Version 1 Rev 0
HARQ Performance Enhancement Techniques

The figure reviews the application of different HARQ types and bit-mapping techniques. The boxes represent the area of applicability. Color intensity increases with achievable gain. In the initial transmission with 16-QAM different bit reliabilities cannot be averaged out. The bit distribution technique (HARQ bit collection) improves the performance of the initial transmission by mapping systematic bits to high reliable positions and parity bits to low reliable bit positions as much as possible. IR allows to benefit from substantial decoding gain, in particular if the initial code rate is high. If the soft combining buffer is limited, chase combining can be applied instead. For 16-QAM, the performance of retransmission can be improved by averaging the bit reliabilities using constellation rearrangement, which works best with chase combining or low initial code rates. For high code rates e.g. > 1/2 full IR is more effective in terms of retransmission as it is quicker in retransmitting all parity bits compared to partial IR. If the UE reports buffer shortage, than only chase combining can be used as it is less memory intensive. Constellation rearrangement gain is only achieved after retransmission and provides increasing advantages with increasing similarity of the transmitted bits in all transmissions (e.g. chase combining or IR with R ~ 1/3).
3-60
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-61
Version 1 Rev 0
HS-PDSCH Interleaving
The interleaving for HSDPA is performed as shown in the figure, separately for each physical channel. For QPSK the interleaver is the same as Rel-99 2nd interleaver. The interleaver is of fixed size: R2=32 rows and C2=30 columns. For 16QAM, there are two identical interleavers of the same fixed size R2C2 = 3230. The output bits from the physical channel segmentation are divided between the interleavers: all odd numbered bits to interleaver one ( upper Interleaver) and all even numbered bits to interleaver two (lower interleaver). Note: The outputs of the interleavers will result in mapping to 16QAM symbols such that the output of first interleaver is mapped to the more reliable positions (i1 and q1) whereas the output of the second interleaver is mapped to the less reliable positions (i2 and q2).
DTX Indication Bits

Note: DTX insertion is not employed. Since only one transport channel per TTI is supported, the rate-matching algorithm is used to fill the available physical resource, instead of using DTX insertion.
3-62
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
3-63
Version 1 Rev 0
3-64
MOTOROLA LTD.2002
HSDPA Protocol Enhancements and Extensions
Version 1 Rev 0
Chapter 4 HSDPA Protocol Enhancements and Extensions
MOTOROLA LTD.2002
4-1
Version 1 Rev 0
HSDPA Protocol Enhancements and Extensions
4-2
MOTOROLA LTD.2002
Objectives
Version 1 Rev 0
Objectives
After this Lecture the Student will be able to: State the interworking between MAC-hs in NodeB and MAC-d in RNC Describe the various functions and components of MAC-hs State the new HS-DSCH FP Control and Data Frame Describe the protocol extensions of NBAP State the basic HSDPA Data Transfer
MOTOROLA LTD.2002
4-3
Version 1 Rev 0
HSDPA Architecture Evolution from Rel 99 / Rel 4

Protocol Structure
The HS-DSCH functionality should be able to operate in an environment where certain cellsdo not support HS-DSCH functionality. The PDCP, RLC and MAC-d layers are almost unchanged from the Rel 99 and Rel 4 architecture. MAC-d is retained in the SRNC. Transport channel type switching is therefore feasible. MAC-d is slightly modified with a link to MAC-hs entity. CRNC may either perform flow control through MAC-c/sh or is transparent depending on UTRAN implementation. Note: The RLC is nearly unchanged and can operate in either AM or UM mode but not in TM mode due to ciphering. PDCP can be configured either to perform or not to perform header compression. The new functionalities of HARQ and HS-DSCH scheduling are included in the MAC layer. In the UTRAN these functions are included in MAC-hs located in Node B. The transport channel that the HS-DSCH functionality uses is called HS-DSCH (High Speed Downlink Shared Channel) and is controlled by the MAC-hs.
HS-DSCH Characteristics
The HS-DSCH transport channel has the following characteristics: An HS-DSCH transport channel is processed and decoded from one CCTrCH; There is only one CCTrCH of HS-DSCH type per UE; The CCTrCH can be mapped to one or several physical channels; There is only one HS-DSCH per CCTrCH; Existence in downlink only; Possibility to use beam forming; Possibility of applying link adaptation techniques other than power control; Possibility to be broadcast in the entire cell; Always associated with a DPCH and one or more shared physical control channels (HS-SCCHs).
4-4
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-5
Version 1 Rev 0
NBAP - HSDPA Message and IE Overview

The table provides an quick overview ( not exhaustive) of the most important NBAP messages and the HSDPA specific IEs. RL Reconfiguration Request and Response relate to unsynchronized radio link change whereas RL Reconfiguration Prepare and Ready relate to synchronized radio link change. The unsynchronized RL Reconfiguration procedure is used when there is no need to synchronize the time of the switching from the old to the new configuration in one Node B used for a UE-UTRAN connection with any other Node B also used for the UEUTRAN connection.
HS-DSCH Resources Information

The Node B shall include the HS-DSCH Resources Information IE for every Cell which has been configured with HS-DSCH resources.
HS-DSCHs MAC-d Flow To Rearrange

This procedure is started by the Node B when bearers for the Node B Communication Context need to be rearranged. The NodeB indicates the MAC-d flow ID for whose bearer shall be rearranged. The Node B may initiate the Bearer Rearrangement procedure at any time after establishing a Radio Link.
HS-DSCH FDD Information

The HS-DSCH FDD Information IE is used for initial addition of HS-DSCH information to a Node B Communication Context.
HS-DSCH MAC-d Flows To Add

If the any HS-DSCH MAC-d Flows To Add or HS-DSCH MAC-d Flows To Delete IEs are included, then the Node B shall use this information to add/delete the indicated HS-DSCH MAC-d flows. When an HS-DSCH MAC-d flow is deleted, all its associated Priority Queues shall also be removed.
HS-DSCH Information To Modify

If the MAC-hs Guaranteed Bit Rate IE is included in the HS-DSCH Information To Modify IE, the Node B shall use this information to optimize MAC-hs scheduling decisions for the related HSDPA Priority Queue. If the MAC-hs Window Size IE or T1 IE is/are included in the HS-DSCH Information To Modify IE, then the Node B shall use the indicated values in the new configuration for the related HSDPA priority queue. If the HS-SCCH Power Offset IE is included in the HS-DSCH Information To Modify IE, the Node B may use this value to determine the HS-SCCH power. The HS-SCCH Power Offset should be applied for any HS-SCCH transmission to this UE. If the Measurement Power Offset IE is included in the HS-DSCH Information IE or in the HS-DSCH Information To Modify IE, then the Node B shall use the measurement power offset as described in chapter 2. If the HS-DSCH Information To Modify IE includes the HS-SCCH Code Change Grant IE, then the Node B may modify the HS-SCCH codes corresponding to the HS-DSCH. The Node B shall then report the codes which are used in the new configuration specified in the HS-SCCH Specific Information Response IE in the RADIO LINK RECONFIGURATION READY message.
HS-DSCH MAC-d Flows To Delete

The HS-DSCH MAC-d Flows To Delete IE is used for the removal of HS-DSCH MAC-d flows from a Node B Communication Context. When an HS-DSCH MAC-d flow is deleted, all its associated Priority Queues shall also be removed.
4-6
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-7
Version 1 Rev 0
NBAP - HSDPA Common Measurement Report

The NodeB can be configured to report common measurements via IE Common Measurement Value Information. Among other measurements, HS-DSCH Requirement Power and Bit Rate are reported within the HS-DSCH Required Power Value Information IE: The Common Measurement Value reports the most recent value for the respectively measurement, for which the reporting criteria were met.
HS-DSCH Required Power Value Information

The HS-DSCH Required Power Value Information IE reports the HS-DSCH Required Power Value IE for each priority class. For each priority class, a list of UEs, identified by the CRNC Communication Context IEs, requiring a particularly high amount of power to meet the Guaranteed Bit Rate for their established HS-DSCH connections may be included. Additionally, the HS-DSCH Required Power Per UE Weight IE may be included for each of those UEs. Scheduling Priority Indicator Indicates the relative priority of the HS-DSCH data frame. Used by the Node B when scheduling HS-DSCH. "0" = Lowest Priority"15" = Highest Priority HS-DSCH Required Power Value The HS-DSCH Required Power Value IE indicates the minimum necessary power for a given priority class to meet the Guaranteed Bit Rate for all the established HS-DSCH connections belonging to this priority class.
HS-DSCH Required Power Per UE Information

List of UEs with Guaranteed Bit Rate indicating their required power consumption relative to the HS-DSCH Required Power Value HS-DSCH Required Power Per UE Weight Expressed in percentage of the value provided in the HS-DSCH Required Power Value IE HS-DSCH Provided Bit Rate Value Information The HS-DSCH Provided Bit Rate Value Information IE reports the HS-DSCH Provided Bit Rate Value IE for each priority class. HS-DSCH Provided Bit Rate Value The HS-DSCH Provided Bit Rate Value IE indicates the HS-DSCH Provided Bit Rate as defined for MAC-hs on page HS-DSCH Bit Rate Measurement Note: MaxNoOfContextsOnUeList means Maximum number of Communication Contexts to include in the list of UEs. MaxNoOfPriorityClasses means Maximum number of HS-DSCH Scheduling Priorities.
4-8
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-9
Version 1 Rev 0
NBAP HSDPA Radio Link Parameter Update Indication

The Radio Link Parameter Update procedure is executed by the Node B when the update of HS-DSCH related radio link parameter values are needed on the Node B side. With this procedure, Node B can suggest some HS-DSCH related Radio Link Parameter values to RNC. The Node B initiates the Radio Link Parameter Update procedure by sending the RADIO LINK PARAMETER UPDATE INDICATION message to the CRNC. The message contains suggested value(s) of the HS-DSCH related parameter(s) that should be reconfigured on the radio link. RADIO LINK PARAMETER UPDATE INDICATION is sent if the NodeB needs to update HS-DSCH related parameters, to allocate new HS-SCCH Codes; here the NodeB shall include HS-SCCH Code Change Indicator IE. update the CQI Feedback Cycle k, CQI Repetition Factor, ACK-NACK Repetition Factor, CQI Power Offset, ACK Power Offset and/or NACK Power Offset, there the Node shall include CQI Feedback Cycle k IE, CQI Repetition Factor IE, ACK-NACK Repetition Factor IE, CQI Power Offset IE, ACK Power Offset IE and/or NACK Power Offset IE
4-10
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-11
Version 1 Rev 0
NBAP - HS-DSCH Related IEs (1)

Allocation / Retention Priority
The NodeB is informed about UEs ARP (Allocation / Retention Priority) which determines the bearer priority relative to other UMTS bearers of the same UE and to other UEs. ARP Value Meaning 0 = not available 1 = highest priority 14 = lowest priority 15 = no priority [3GTS 25.433 (9.2)]
Priority Flow Information

Scheduling Priority Indicator This parameter is set by CRNC when the MAC-d flow(s) are to be established or modified. It is used by the packet scheduler in NodeB to prioritise MAC-d flows relative to others at Uu. Discard Timer The Discard Timer IE defines the time to live for a MAC-d PDUs (MAC-hs SDU) starting from the instant of its arrival into an HSDPA Priority Queue. The Node B shall use this information to discard out-of-data MAC-hs SDUs from the HSDPA Priority Queues. MAC-hs Guaranteed Bit Rate The MAC-hs Guaranteed Bit Rate IE indicates the guaranteed number of bits per second that Node B should deliver over the air interface under normal operating conditions (provided there is data to deliver).
4-12
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-13
Version 1 Rev 0

MAC-d PDU Size
RLC Mode The RLC Mode IE indicates the RLC Mode used for a Priority Queue.
UE Capabilities Information
HS-SCCH Power Offset The HS-SCCH Power Offset IE indicates the Power offset relative to the pilot bits on the DL DPCCH.
4-14
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-15
Version 1 Rev 0
NBAP HS-DSCH Info Response

The HS-DSCH Information Response provides information for HS-DSCH that have been established or modified. It also provides additional HS-DSCH information determined within the Node B
HS-DSCH Initial Capacity Allocation

For every MAC-d flow being established, the NodeB shall indicate the Initial Capacity Allocation IE if the NodeB allows CRNC to forward MAC-d PDUs before the NodeB has allocated capacity on the user plane via HS-SCCH. The HS-DSCH Initial Capacity Allocation IE provides flow control information for each scheduling priority class for the HS-DSCH FP over Iub. Indicates the initial number of MAC-d PDUs that may be transmitted before new credits are received from the Node B. Note that the initial capacity allocation is only valid for the first HS-DSCH DATA FRAME. After that the NodeB should send a Capacity Allocation control frame towards CRNC declaring how many MAC-d PDUs can be transmitted in what time interval at Uu for the respective UE.
HS-SCCH Specific Information Response

The Node B shall allocate HS-SCCH codes corresponding to the HS-DSCH and include the HS-SCCH Specific Information Response IE.
HARQ Memory Partitioning Information

The Node B shall include the HARQ Memory Partitioning IE in the HS-DSCH Information Response IE.
4-16
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-17
Version 1 Rev 0
HS-PDSCH and HS-SCCH Power and Code Information
This page intentionally left blank
4-18
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-19
Version 1 Rev 0
HSDPA Cell Configuration - AUDIT
4-20
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-21
Version 1 Rev 0
HSDPA Data Transfer Procedure (1)
4-22
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-23
Version 1 Rev 0
4-24
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-25
Version 1 Rev 0
RRC Protocol Enhancement New IEs with HSDPA

The figure summarizes the most important RRC messages (not exhaustive) which allow to configure or reconfigure HSDPA configuration in UE. At the same time all the important HSDPA relevant IEs and group names are depicted which will be explained in detail on the following pages.
MP Mandatory Present
A value for that information is always needed, and no information is provided about a particular default value. If ever the transfer syntax allows absence (e.g., due to extension), then absence leads to an error diagnosis.
MD Mandatory Default
A value for that information is always needed, and a particular default value is mentioned (in the semantical information column). This opens the possibility for the transfer syntax to use absence or a special pattern to encode the default value. Example: Activation Time If there is no value for Activation Time then the mandatory default value is now.
CV Conditional Value
The need for a value for that information depends on the value of some other IE or IEs, and/or on the message flow (e.g., channel, SAP). The need is specified by means of a condition, the result of which may be that the information is mandatory present, mandatory with default value, not needed or optional. If one of the results of the condition is that the information is mandatory present, the transfer syntax must allow for the presence of the information. If in this case the information is absent an error is diagnosed. If one of the results of the condition is that the information is mandatory with default value, and a particular default value is mentioned (in the Semantical information column), the transfer syntax may use absence or a special pattern to encode the default value. If one of the results of the condition is that the information is not needed, the transfer syntax must allow encoding the absence. If in this case the information is present, it will be ignored. In specific cases however, an error may be diagnosed instead. If one of the results of the condition is that the information is optional, the transfer syntax must allow for the presence of the information. In this case, neither absence nor presence of the information leads to an error diagnosis.
OP - Optional
The presence or absence is significant and modifies the behavior of the receiver. However whether the information is present or not does not lead to an error diagnosis.
4-26
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-27
Version 1 Rev 0
Added or Reconfigured MAC-d Flow

MAC-d Flow Identity
If the IE "Added or Reconfigured MAC-d flow" is included, the UE shall: For each MAC-hs queue included in the IE "MAC-hs queue to add or reconfigure" list: delete the old mapping if the UE has previously stored a mapping between this MAC-hs queue and another (previous) MAC-d flow map the MAC-d flow indicated in the current message to this MAC-hs queue; set the release timer T1 for each of the MAC-hs queues in the MAC-hs entity to the value in the corresponding IE "T1"; set the MAC-hs receiver window size for each of the MAC-hs queues in the MAC-hs entity to the value in the corresponding IE "MAC-hs window size"; configure MAC-hs with the mapping between MAC-d PDU sizes index and allowed MAC-d PDU sizes as follows: for each MAC-hs queue included in the IE "MAC-hs queue to delete" list: delete any information about the MAC-hs queue identified by the IE "MAC-hs queue Id" If a DL MAC-d flow is requested to be deleted: delete any information about the DL HS-DSCH MAC-d flow identified by the IE "MAC-d Flow Identity", i.e. delete any information about MAC-hs queue(s) mapped onto this MAC-d flow.
MAC-hs Queue Id
The MAC-hs queue ID is unique across all MAC-d flows.
T1 Reorder Release Timer

Timer governing the release of MAC-d PDUs carried by MAC-hs PDUs to the upper layers even though there are outstanding MAC-hs PDUs with lower TSN values.
MAC-d PDU Size Info

Mapping of the different MAC-d PDU sizes configured for the HS-DSCH to the MAC-d PDU size index in the MAC-hs header.
4-28
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-29
Version 1 Rev 0
Uplink DPCH Power Control Info and Others

Uplink DPCH Power Control Info
If the IE "Downlink HS-PDSCH Information" is included and the UE would enter CELL_DCH state, the UE shall: if the UE has received IE "Uplink DPCH Power Control Info" and stored Ack, Nack and Ack-Nack Repetition factor; and if the UE has stored IEs "MAC-hs queue to add or reconfigure list", "MAC-d PDU size Info" and "RB Mapping Info" corresponding to the HS-PDSCH configuration; and if the UE has stored IEs "MAC-hs queue to add or reconfigure list", "MAC-d PDU size Info" and "RB Mapping Info" corresponding to the HS-PDSCH configuration; set the variable HS_DSCH_RECEPTION to TRUE; start HS-DSCH reception procedures according to the stored HS-PDSCH configuration: if the IE "New H-RNTI" is included: use the value of the variable H_RNTI as UE identity in the HS-SCCH reception procedure in the physical layer. If the IE "HS-SCCH Info" is included, the UE shall: receive the HS-SCCH(s) according to the IE "HS-SCCH channelization code" on the serving HS-DSCH radio link applying the scrambling code as received in the IE "DL Scrambling code". (mandatory default is primary scrambling code) if the IE "Measurement Feedback Info" is included: use the information for the CQI procedure in the physical layer on the serving HS-DSCH radio link.
MAC-hs Reset
After a MAC-hs reset; among other things (which are explained on page UE MAC-hs Other Functions); indicate to all AM RLC entities mapped on HS-DSCH to generate a status report. Note: The UE shall generate a status report, even if the STATUS prohibit or the EPC (estimated PDU counter) mechanism is configured to prohibit the Receiver from sending a status report containing any of the SUFIs LIST, BITMAP, RLIST or ACK if an indicator is received from MAC requesting generation of a status report following MAC-hs reset.
Serving HS-DSCH Radio Link Indicator

If the IE "Serving HS-DSCH radio link indicator" is set to TRUE consider this radio link as the serving HS-DSCH radio link.
4-30
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-31
Version 1 Rev 0
MAC-hs Protocol
MAC-hs Protocol
The MAC protocol in the RNC will decide on which transport channel the MAC-d PDU will be transmitted. For the HS-DSCH transport, the MAC is further split into the NodeB MAC called the MAC-hs (MAC high speed), which enables fast radio resource allocation. The MAC-hs takes care of the transport block scheduling, HARQ channel allocation and transport format selection. The forward signaling on the HS-SCCH serves the MAC-hs at the receiver and contains, beside terminal identification, parameters for TFRC (number and indexes of the channelization codes and modulation), HARQ process identifier, TB size, redundancy version and new data indicator. The reverse signaling on the HS-DPCCH contains the CQI-report and the HARQ acknowledgements. The presented figure is only informative which aims to provide an overall picture. Note two MAC protocol configurations are possible on the UTRAN side. As already mentioned in chapter 1, configuration with or without MAC-c/sh is implementation dependent (vendor specific).
MAC-hs - co-incident CRNC and SRNC

Figure here shows the protocol model for the HS-DSCH transport channel managed by MAC-hs when the Controlling and Serving RNC are co-incident. The MAC-hs entity in the Node B transfers MAC-hs PDU to the peer MAC-hs entity in the UE over the Uu interface. The MAC-d entity in the RNC transfers MAC-d PDUs to the MAC-hs in the Node B using the services of the HS-DSCH FP entity (HS-DSCH frame protocol). The HS-DSCH FP entity adds header information to form a HS-DSCH FP PDU that is transported to the Node B over a transport bearer which could be based on IP or ATM. A relaying function in the Node B relays the HS-DSCH frame received by HS-DSCH FP entity to the MAC-hs entity. HS-DSCH scheduling is performed by MAC-hs in the Node B. Note: HS-DSCH is applicable only to PS domain RABs.
HS-DSCH RNTI
The HS-DSCH RNTI (=H-RNTI) is used for the UE specific CRC attachment in HS-SCCH. HS-DSCH RNTI is allocated by controlling RNC upon UE establishing a HS-DSCH channel. HS-DSCH RNTI shall be unique within the cell carrying the HS-DSCH.
4-32
MOTOROLA LTD.2002
MAC-hs Protocol
Version 1 Rev 0
MAC-hs Protocol
MOTOROLA LTD.2002
4-33
Version 1 Rev 0
MAC-hs Tasks
MAC-hs Tasks
The MAC-hs is responsible for handling the data transmitted on the HS-PDSCH. Furthermore it is responsible for the management of the physical resources allocated to HS-PDSCH. MAC-hs receives configuration parameters from the RRC layer via the MAC-Control SAP. There is priority handling per MAC-d PDU in the MAC-hs. MAC-hs has various tasks as outlined below:
Iub Flow Control

A new flow control function is included to support the data transfer between MAC-d and MAC-hs. In case of MAC-c/sh, depending on UTRAN implementation, there is a flow control function between MAC-hs and MAC-c/sh. The flow control provides for a controlled data flow between the entities taking the transmission capabilities of the air interface into account in a dynamic manner. This flow control function should limit layer 2 signaling latency and reduce discarded and retransmitted data in case of HS-DSCH congestion or cell change.
Packet Buffering and Scheduling

This function manages HS-DSCH resources between HARQ entities and data flows according to their priority. Based on status reports from associated uplink signaling either new transmission or retransmission is determined. The HS-DSCH TB must be buffered until the acknowledgement has been received via HS-DPCCH.
TFRC Selection
TFRC (transport format and resource combination) is the selection of an appropriate transport format and resource combination for the data to be transmitted on HS-DSCH. The TFRI is the indicator for the TFRC selection. The TFRI is explicitly signaled on HS-SC.CH which indicates the TB size, channelization code set and modulation scheme. Compared to that the MAC-d PDU size is implicitly signaled via MAC-hs PDU header.
AMC
As explained in chapter 2 and 3 already AMC function provides for the initial MCS selection based on CQI reporting, DL PC, Ack/Nack ratio and UE capability.
HARQ
The HARQ protocol is based on an asynchronous downlink (N-SAW scheme) and synchronous uplink scheme (CQI reporting). The ARQ combining scheme is based on Incremental redundancy. Chase combining is considered to be a particular case of Incremental Redundancy. The UE soft memory capability is defined according to the needs for chase combining. The soft memory is partitioned across the HARQ processes in a semi-static fashion through upper layer signaling. The UTRAN should take into account the UE soft memory capability when configuring the different transport formats (including possibly multiple redundancy versions for the same effective code rate) and when selecting transport formats for transmission and retransmission. Ack / Nack and CQI Reporting is performed by UE using HS-DPCCH.
4-34
MOTOROLA LTD.2002
MAC-hs Tasks
Version 1 Rev 0
MAC-hs Tasks
MOTOROLA LTD.2002
4-35
Version 1 Rev 0
HS-DSCH MAC PDU
HS-DSCH MAC PDU

General
A MAC PDU is a bit string with a length not necessarily a multiple of 8 bits. MAC SDUs, depending on the provided service, are bit strings with any non-null length or bit strings with an integer number of octets in length. An SDU is included into a MAC PDU from first bit onward.
MAC-hs SDU
For HS-DSCH the MAC-d PDU format equals the MAC PDU format for non HS-DSCH case. The MAC-d PDU header for DTCH and DCCH mapped on HS-DSCH includes the C/T field if multiplexing on MAC is applied. In case of HS-DSCH a MAC PDU consists of one MAC-hs header and one or more MAC-hs SDUs where each MAC-hs SDU equals a MAC-d PDU. A maximum of one MAC-hs PDU can be transmitted in a TTI per UE. The MAC-hs header is of variable size. The MAC-hs SDUs in one TTI belongs to the same reordering queue. There need to be per PDP context / QoS profile one (re-)ordering queue configured in NodeB and UE in order to allow the individual user bearer(s) (RLC instances) being served by HSDPA according to their QoS needs.
MAC-hs Header of MAC-hs PDU

VF (version flag) is a one bit flag providing extension capabilities of the MAC-hs PDU format. The VF field is set to zero and the value one is reserved in this version of the protocol. The Queue ID field provides identification of the reordering queue in the receiver, in order to support independent buffer handling of data belonging to different reordering queues. The length of the Queue ID field is 3 bit. TSN (transmission sequence number) field provides an identifier for the transmission sequence number on the HS-DSCH. The TSN field is used for reordering purposes to support in-sequence delivery towards MAC-d layer in UE. The SID (size index identifier) field identifies the size of a set of consecutive MAC-d PDUs. The MAC-d PDU size for a given SID is configured by NBAP for NodeB (RRC for UE) and is independent for each Queue ID. N field identifies the number of consecutive MAC-d PDUs with equal size. The length of the N field is 7 bits. In FDD mode, the maximum number of PDUs transmitted in a single TTI shall be assumed to be 70. If more PDUs than the defined maximum number of PDUs for the corresponding mode are received, the UE behavior is unspecified. The F field is a flag indicating if more fields are present in the MAC-hs header or not. If the F field is set to "0" the F field is followed by an additional set of SID, N and F fields. If the F field is set to "1" the F field is followed by a MAC-d PDU. The maximum number of MAC-hs header extensions, i.e. number of fields F set to 0, in a single TTI shall be assumed to be 7. If more extensions than the maximum defined for the corresponding mode are included in a TTI, the UE behavior is unspecified.
4-36
MOTOROLA LTD.2002
HS-DSCH MAC PDU
Version 1 Rev 0
HS-DSCH MAC PDU
MOTOROLA LTD.2002
4-37
Version 1 Rev 0
MAC-hs Entity UTRAN Side

There is one MAC-hs entity in the UTRAN for each cell that supports HS-DSCH transmission.
Scheduling/Priority Handling
This function manages HS-DSCH resources between HARQ entities and data flows according to their priority. Based on status reports from HS-DPCCH either new transmission or retransmission is determined. Further it determines the Queue ID and TSN for each new MAC-hs PDU being serviced. To maintain proper transmission priority, a new transmission can be initiated instead of a pending retransmission at any time to support the priority handling. The TSN is unique to each priority class within a HS-DSCH, and is incremented for each new data block. This means every queue has its own priority class value determined by the PDP context / QoS profile. The priority class per priority queue is assigned through NBAP. The PS may change the queues priority dynamically in order to manage various PDP contexts one user and also among different users. The queues priority determines when it is going to be serviced, means MAC-hs PDUs are transmitted on HS-DSCH. Note: It is not permitted to schedule new transmissions, including retransmission originating in the RLC layer, within the same TTI, along with retransmission originating from the HARQ layer.
HARQ
One HARQ entity handles the hybrid ARQ functionality for one user. One HARQ entity is capable of supporting multiple instances (HARQ process) of stop and wait HARQ protocols. There is one HARQ process per HS-DSCH per TTI.
TFRC Selection
Selection of an appropriate transport format and resource for the data to be transmitted on HS-DSCH. Note: the TFRC selection can be compared with the TFC selection on legacy UMTS transport channels.
MAC-d / Flow Control

The MAC-d PDUs to be transmitted are transferred via Iub/Iur interface in case of configuration without MAC-c/sh. In case with MAC-c/sh the MAC-d PDUs are transferred via the Iub interface from MAC-c/sh to the MAC-hs. A flow control function exists between MAC-hs and MAC-d to limit buffering. In case with MAC-c/sh, a flow control exists between MAC-c/sh and MAC-hs. Note: Flow control is provided independently per priority class for each MAC-d flow.
4-38
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-39
Version 1 Rev 0
MAC-hs Entity UE Side

The MAC-hs handles the HSDPA specific functions. In the pictured model the MAC-hs comprises the following entities:
HARQ
The HARQ entity is responsible for handling the MAC functions relating to the HARQ protocol. The HARQ functional entity handles all the tasks that are required for hybrid ARQ. It is responsible for generating Acks or Nacks. The detailed configuration of the hybrid ARQ protocol is provided by RRC over the MAC-Control SAP. During HSDPA data reception, the UE obtains the HARQ (HARQ process ID, NDI bit and X(RV)) and TFRI information from the HS-SCCH. HARQ and TFRI info provide the necessary input parameter for the HARQ protocol.
Reordering Queue distribution

The reordering queue distribution function routes the MAC-hs PDUs to the correct reordering buffer based on the Queue ID. Note: Per PDP context there should be a queue configured in order to allow handling per QoS profile. More than one PDP context could be necessary if applications have different QoS requirements. Then also more than one queue needs to be configured. The re-ordering queue distribution function avoids duplicate of MAC-hs PDUs based on their TSNs.
Reordering
The reordering entity reorders received MAC-hs PDUs according to their TSN. MAC-hs PDUs with consecutive TSNs are delivered to the disassembly function upon reception. MAC-hs PDUs are not delivered to the disassembly function if MAC-hs PDUs with lower TSN are still missing. There is one reordering entity for each Queue ID configured at the UE.
Disassembly
The disassembly entity is responsible for the disassembly of MAC-hs PDUs. When a MAC-hs PDU is disassembled the MAC-hs header is removed, the MAC-d PDUs are extracted and any present padding bits are removed. Then the MAC-d PDUs are delivered to higher layer.
MAC-d
The MAC-d entity has a connection to the MAC-hs entity. This connection is used to receive data from the HS-DSCH transport channel which is handled by MAC-hs (downlink). There is one MAC-d entity in the UE.
4-40
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-41
Version 1 Rev 0
HS-DSCH MAC Architecture UTRAN Side

This subclause describes the modification to the MAC model with respect to the Rel. 99 model to support the features for HS-DSCH on the UTRAN side.
Overall Architecture
The MAC-hs is added to the MAC architecture of Rel. 99. The MAC-hs is located in the NodeB. If an HS-DSCH is assigned to the UE, the MAC-hs SDUs, i.e. the MAC-d PDUs to be transmitted are transferred from MAC-c/sh to the MAC-hs via the Iur/Iub interface in case of configuration with MAC-c/sh, or from the MAC-d via Iur/Iub in case of configuration without MAC-c/sh. The HS-DSCH FP is used to transfer the MAC-d PDUs from SRNC towards MAC-hs in NodeB by using the HS-DSCH Data Frame.
MAC-c/sh (optional)
The data for the HS-DSCH is subject to flow control between the SRNC and the DRNC/CRNC. Therefore a new flow control function is included in DRNC/CRNC to support the data transfer between MAC-d and MAC-hs.
MAC-d Flow in UTRAN

The figure shall illustrate how the MAC-d flow are mapped to priority queues in MAC-hs. We depicted an example for one UE which could have three MAC-d flows. (One could also interpret the figure as an example of handling three HSDPA UEs each with their own MAC-d flow independently from each other. In this case every MAC-d flow runs in a different VPI/VCI/CID.) MAC-d flow: 0 If more than one DTCH is mapped onto the same MAC-d flow the C/T field is necessary in order to distinguish the different logical channels. If the individual DTCHs within the same MAC-d flow have different priority then they are assigned each to a certain priority queue. MAC-hs allows to associate a scheduling priority indicator to each logical channel within the same MAC-d flow. Therefore the individual DTCH is distinguished by CmCH-PI (common TrCH priority indicator) which corresponds to the scheduling priority indicator in MAC-hs. The MAC-hs (priority) queue ID is unique across all MAC-d flows. In our example we chose priority values from 4 to 7 indicating that a MAC-d flow can be mapped to one or more priority queues if the DTCHs are associated with different QoS profiles. MAC-d flow: 1 One DTCH is mapped onto one MAC-d flow associated to one priority queue ID. MAC-d flow: 2 Two DTCHs with the same QoS profile are mapped onto the same MAC-d flow and therefore also to the same priority queue.
4-42
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-43
Version 1 Rev 0
HS-DSCH Architecture UE Side

This subclause describes the architecture of the MAC and functional split required to support HS-DSCH on the UE side.
Overall Architecture
The figure shows the overall MAC architecture up to RLC level. The data received on the HS-DSCH is mapped to the MAC-hs. The MAC-hs is configured via RRC signaling to set the parameters in MAC-hs such as allowed transport format combinations for the HS-DSCH. The associated downlink signaling on HS-SCCH carries information for support of HS-DSCH while the associated uplink signaling on HS-DPCCH carries feedback information.
MAC-d Flow in UE
The figure shows the MAC-d flow mapping of one UE. The MAC-d entity is modified with the addition of a link to the MAC-hs entity. Note that the links to MAC-hs and MAC-c/sh cannot be configured simultaneously in one UE. This means that the UE is not able to support HS-DSCH and additional downlink shared or common TrCHs beside DCH. The mapping between C/T MUX entity in MAC-d and the reordering buffer in MAC-hs is configured by RRC signaling. One reordering buffer maps to one C/T MUX entity but many reordering buffer can map to the same C/T MUX entity. Note: De-/Ciphering is only performed for RLC-TM in MAC-d, for RLC-UM and RLC-AM de-/ciphering is performed in RLC. This means that RLC-TM is not supported on HS-DSCH as no frame number exists for HS-DSCH to count the 2 ms subframe continuously. Reordering / Disassembly The reordering entity organizes received HS-DSCH data blocks (MAC-hs PDUs) according to their TSN. MAC-hs PDUs with consecutive TSNs are delivered to higher layers upon reception. One MAC-hs PDU contains only MAC-d PDUs with the same priority and from the same MAC-d flow. Different MAC-d PDU sizes can be supported in a given MAC-hs PDU due to the size index parameter SID. MAC-d flow: 0, 1, 2 One or more queues can be associated with the same MAC-d flow. The MAC-hs queue ID is unique across all MAC-d flows. Thus the queue ID within MAC-hs PDU header identifies the associated MAC-d flow. If more than one DTCH is mapped onto the same MAC-d flow, the C/T field is necessary. A MAC-d flow can be associated with more than one queue. In such a case, the queues must have different IDs.
4-44
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-45
Version 1 Rev 0
HS-DSCH Scheduler Functions

The scheduler in NodeB performs following functions:
UE Scheduling
HS-DSCH scheduler schedules all UEs within a cell.
Serving Priority Queues

The scheduler schedules MAC-hs SDUs based on information through NBAP signaling. One UE may be associated with one or more MAC-d flows. Each MAC-d flow contains HS-DSCH MAC-d PDUs for one or more priority queues. This means one or more PDP context / QoS Profile. The scheduler also determines the HARQ entity and the queue to be serviced.
Updating MAC-hs Header

The HS-DSCH scheduler sets the TSN for new data blocks being transferred from the selected queue. It sets the TSN to value 0 for the first MAC-hs PDU transmitted for each Queue ID within an HS-DSCH. It increments the TSN with one for each transmitted MAC-hs PDU on each Queue ID within an HS-DSCH. NOTE: The scheduler may re-use TSNs by toggling the NDI bit in order to resume pre-empted transmissions or to force the UE to flush the soft buffer. In this case the content of the payload may be changed but care should be taken to preserve the higher layer data order. Indicates the Queue ID and TSN to the HARQ entity for each MAC-hs PDU to be transmitted.
Scheduling new Transmission and retransmission

Based on the status reports from HARQ processes the scheduler determines if either a new transmission or a retransmission should be made. A new transmission can however be initiated on a HARQ process at any time. Based on a delay attribute provided by upper layers (RRC), the scheduler may decide to discard any out-of-date MAC-hs SDU.
Determining RV
The scheduler determines a suitable redundancy version for each transmitted and retransmitted MAC-hs PDU and indicates the redundancy version to lower layer ( HS-SCCH).
4-46
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-47
Version 1 Rev 0
NodeB HARQ
NodeB HARQ
HARQ Entity
There is one HARQ entity per UE in UTRAN. The HARQ entity sets the Queue ID in transmitted MAC-hs PDUs to the value indicated by the UTRAN scheduler; sets the transmission sequence number (TSN) in transmitted MAC-hs PDUs to the value indicated by the UTRAN scheduler; sets the HARQ process identifier in transmitted MAC-hs PDUs; determines a suitable HARQ process to service the MAC-hs PDU and set the HARQ process identifier accordingly.
HARQ Process
The HARQ process sets the NDI (new data indicator) in transmitted MAC-hs PDUs. The HARQ process sets the New Data Indicator to the value "0" for the first MAC-hs PDU transmitted by a HARQ process does not increment the NDI for retransmission of a MAC-hs PDU; increments the NDI with one for each transmitted MAC-hs PDU containing new data; processes received status messages; delivers received status messages e.g. Ack/Nack or DTX to the scheduler which prepares for new or retransmission<.
Error Handling in HARQ

The most frequent error cases to be handled are the following: NACK is detected as an ACK. The MAC-h starts afresh with new data in the HARQ process. The data block is discarded by MAC-hs from the buffer and lost. Retransmission is left up to higher layers, e.g. RLC. ACK is detected as a NACK: If the MAC-hs retransmits the data block, the UE will re-send an ACK to the network. If in this case the MAC-hs in NodeB sends an abort indicator by incrementing the NDI, the receiver at the UE will continue to process the data block as in the normal case. If a CRC error on the HS-SCCH is detected, UE receives no data and sends no status report via HS-DPCCH. If the absence of the status report is detected, MAC-hs can retransmit the block. Even RLC may retransmit in case of extensive time-out means continuous HS-SCCH decoding failure. Note: If the first transmission of a HS-DSCH data block fails by detecting DTX in uplink on the designated HS-DPCCH subframe instead of Ack/Nack the NodeB should not retransmit with Full IR scheme as obviously nothing has been received by the UE in the first place HS-SCCH decoding failure.
4-48
MOTOROLA LTD.2002
NodeB HARQ
Version 1 Rev 0
NodeB HARQ
MOTOROLA LTD.2002
4-49
Version 1 Rev 0
HS-DSCH Bit Rate Measurement

The bit rate measurements provided by HS-DSCH is defined as follows: for each priority class the MAC-hs entity measures the total number of MAC-d PDU bits MAC-hs in NodeB considers only the successful MAC-d PDUs ( Ack received via HS-DPCCH) during the last measurement period, divided by the duration of the measurement period; the values reported by MAC-hs shall be raw samples; the measurement period shall be [100 ms]. The HS-DSCH Provided Bit Rate Value Information IE on NBAP reports the HS-DSCH Provided Bit Rate Value for each priority class within the Common Measurement Report. Note: If the RADIO LINK SETUP REQUEST or RADIO LINK RECONFIGURATION PREPARE or RADIO LINK RECONFIGURATION REQUEST message includes the MAC-hs Guaranteed Bit Rate IE for a Priority Queue in the HS-DSCH MAC-d Flows Information IE in the HS-DSCH Information IE, then the Node B shall use this information to optimize MAC-hs scheduling decisions for the related HSDPA Priority Queue.
4-50
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-51
Version 1 Rev 0
UE MAC-hs Operation
UE MAC-hs Operation
The UE operation in support of HS-DSCH is split among the following functional units: HARQ entity, HARQ process, reordering entity and disassembly entity
UE HARQ
HARQ Entity There is one HARQ entity at the UE which processes the HARQ process identifiers received on the HS-SCCH transmissions associated with MAC-hs PDUs received on the HS-DSCH. A number of parallel HARQ processes are used in the UE to support the HARQ entity. The number of HARQ processes is configured RRC and limited to maximum of eight: Each received MAC-hs PDU shall be allocated to the HARQ process indicated by the HARQ process identifier of the MAC-hs PDU. HARQ Process Discard MAC-hs PDUs The UE may, if the MAC-hs PDU is received within 5 sub-frames from the reception of the previous MAC-hs PDU intended for this HARQ process; discard the MAC-hs PDU. If the NDI (one bit) has been incremented compared to the value in the previous received transmission in the very HARQ process or it is the first received transmission in the HARQ process, the UE shall replace the data being currently in the soft buffer. If the NDI is identical to the value used in the previous received transmission in the HARQ process, combine the received data with the data currently in the soft buffer for this HARQ process. Note: the HARQ process processes the Queue ID in the received MAC-hs PDU. The UE shall arrange the received MAC-hs PDUs in queues based on the Queue ID:
4-52
MOTOROLA LTD.2002
UE MAC-hs Operation
Version 1 Rev 0
UE MAC-hs Operation
MOTOROLA LTD.2002
4-53
Version 1 Rev 0
Transmitter and Receiver Stall Avoidance

The MAC layer needs to guarantee delivery of data in-order to higher layers. Since the HARQ mechanism using multiple HARQ channels may result in data being recovered out of sequence by the UE, a re-ordering sub-layer was added in the MAC layer in W-CDMA Release 5. The re-ordering sub-layer buffers packets as they are recovered, re-orders these packets, and delivers consecutive packets (as determined by their TSNs) to higher layers. If the re-ordering sub-layer detects missing packets, based on gaps or holes in the TSNs of the recovered packets, then it holds off on (i.e., stalls) the delivery of all packets with TSNs later than the TSN of the earliest missing packet. When the missing packets are finally recovered, the re-ordering sub-layer would then provide these newly recovered packets as well as any prior recovered packets that have been stalled. Three "stall avoidance" mechanisms are provided by W-CDMA Release 5 to allow for practical implementations and to avoid situations where the re-ordering entity waits indefinitely for data that is not retransmitted. These stall avoidance mechanisms include:
Timer-Based Scheme
Each time a missing packet stalls the delivery of packets to higher layers at the receiver, a timer T1 is started. If no other missing packets are thereafter detected then, once the long timer expires, the missing packet is assumed to have been lost, and all packets stalled by this missing packet are then delivered to higher layers. This mechanism requires the maintenance of one timer T1 per re-ordering queue (or a maximum of eight long timers for the eight re-ordering queues defined in W-CDMA Release 5).
Window Bases Scheme

Since each packet is tagged with a specific TSN, the recovered packets may be assembled in the proper sequence at the UE. Although the packets may be initially transmitted in sequential order by the Node B, these packets may be recovered out of order because a variable number of retransmissions may be required for each packet. Once the oldest not yet acknowledged TSN x and the latest received TSN y differ by more than the window size, the MAC-hs PDU x is assumed to be lost and falls out of the window. Thus TX/RX window size determines how long a certain MAC-hs PDU can be retransmitted.
HARQ Activity Scheme

The third scheme to avoid stalling the delivery of recovered packets to higher layers is to detect activity on the HARQ channels. When no packets are expected on any of the HARQ channels (i.e., all of the previous packet transactions are completed), then the data in all re-ordering queues can be delivered by the re-ordering entities to higher layers. This mechanism has a few shortcomings. First, this scheme requires that no packet transaction is pending on any of the HARQ channels in order to be able to flush the packets to higher layers. Second, the receiver "marks off" a HARQ channel only if the packet transaction on that channel is completed. Since the receiver may wait indefinitely for a packet to be recovered on a given HARQ channel (e.g., if the transmitter abandons the packet transaction), the re-ordering queues may never be flushed. Third, if a control message is missed (i.e., not detected by the receiver), then it is possible that the packet associated with this control message will be discarded by the re-ordering entity, if it is subsequently recovered. This would be the case if another packet with a later TSN is recovered and provided to the re-ordering entity before the packet with the missed control message is retransmitted and the control channel is successfully decoded.
4-54
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-55
Version 1 Rev 0
UE Reordering Entity
Parameters
TRANSMIT_WINDOW_SIZE is the size of the transmitter window. This is a parameter in the Node B and the value of the parameter is configured through NBAP signaling. The NodeB stops transmitting new MAC-hs PDUs if their TSN is out of the transmitters window size because there are too many unacknowledged MAC-hs PDUs in the transmit buffer. Note that this handling is performed per priority queue. RECEIVE_WINDOW_SIZE is the size of the receiver window. This is a parameter in the UE and the value of the parameter is configured through RRC signaling.
State Variables
All state variables are non-negative integers. MAC-hs PDUs are numbered by modulo integer TSN cycling through the field 0 to 63. All arithmetic operations on next_expected_TSN, RcvWindow_UpperEdge, T1_TSN and TSN_flush are affected by the 64 modulus. When performing arithmetic comparisons of state variables or TSN values a 64 modulus base is used. This modulus base is subtracted (within the appropriate field) from all the values involved and then an absolute comparison is performed. next_expected_TSN is the Transmission sequence number (TSN) following the TSN of the last in-sequence MAC-hs PDU received. It shall be updated upon the delivery to the disassembly entity of the MAC-hs PDU with TSN equal to next_expected_TSN. The initial value of next_expected_TSN =0. RcvWindow_UpperEdge represents the TSN, which is at the upper edge of the receiver window. After the first MAC-hs PDU has been received successfully, it also corresponds to the MAC-hs PDU with the highest TSN of all received MAC-hs PDUs. The initial RcvWindow_UpperEdge equals 63. RcvWindow_UpperEdge is updated based on the reception of new MAC-hs PDU. T1_TSN: The TSN of the latest MAC-hs PDU that cannot be delivered to the disassembly entity, when the timer T1 is started.
Timer T1
Re-ordering release timer T1 controls the stall avoidance in the UE reordering buffer. The value of T1 is configured by upper layers. ( through RRC).
Receiver Window
Receiver window: defines TSNs of those MAC-hs PDUs that can be received in the receiver without causing an advancement of the receiver window. The size of the receiver window equals RECEIVE_WINDOW_SIZE and spans TSNs going from RcvWindow_UpperEdge RECEIVE_WINDOW_SIZE + 1 to RcvWindow_UpperEdge included.
4-56
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-57
Version 1 Rev 0
Timer and / or Window Based Stall Avoidance

Window Based Scheme
The diagram graphically illustrates the window maintained for a particular priority queue. To resolve ambiguity in the TSN number space caused by the fact that the TSN field has a finite size, the receiver uses a window. The max. size of this window is typically set at less than half of the TSN number space (i.e., < 32) and may be set even smaller according to higher layer requirements. Since the window size is smaller than the TSN number space, the order of the packets within the window is unambiguous. In determining the size of the window, there is a tradeoff. If the window is small, the stall avoidance performance at the receiver increases and the receiver buffer size requirement in UE is reduced. However, the stall probability at the transmitter or the probability of needing to interrupt retransmission (depending on the transmission strategy) is increased. The window is advanced forward as new packets are received. For the receiver, the leading edge of the window is set equal to the "latest" TSN of all recovered packets (RcvWindow_UpperEdge). Packets toward the left of the window have successively "earlier" TSNs. Since the TSN value can wrap around, the latest TSN value may actually be smaller than an earlier TSN whenever the TSN wraps around. Missing packets with TSNs earlier than the trailing edge of the window may be assumed to be lost. Thus, as the window is advanced forward, packets earlier than the trailing window edge are "flushed" and sent to higher layers. This window mechanism may therefore be used to flush out missing packets at the receiver. However, since the size of the window would need to be large to allow for a large number of retransmissions, a large amount of data would be needed to flush out the missing packets. Consequently, the window-based scheme is marginally effective at the end of data bursts, which are frequent in the case of bursty closed loop traffic such as what is generated by web browsing.
Timer Bases Scheme

In order to address the limitations of the window-based scheme, a timer-based mechanism was also introduced. For the timer-based scheme, each time a missing packet stalls the delivery of packets to higher layers at the receiver, a timer T1 is started. If no other missing packets are thereafter detected then, once T1 expires, the missing packet is assumed to have been recovered, and all packets stalled by this missing packet are then delivered to higher layers. This mechanism requires the maintenance of one timer per re-ordering queue. To ensure proper HARQ operation, the timer needs to be set longer than the longest amount of time it takes to complete all retransmission for a given packet. A large number of retransmissions may need to be performed to recover the missing packet. Moreover, in a system that asynchronously schedules retransmission and where the amount of resources (e.g., as quantified by channelization codes and transmit power) available for HSDPA can change dynamically, the amount of time it takes to complete all retransmission for a missing packet can vary widely. Consequently, the value of this timer would need to be long. Otherwise, the retransmission for the missing packet may be prematurely terminated by the expiration of the T1, in which case the missing packet would need to be retransmitted by higher layers, which is highly undesirable. The re-ordering entity may need to wait a significant amount of time for the long timer to expire or until all the retransmissions for the missing payload are completed.
4-58
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-59
Version 1 Rev 0
UE MAC-hs Other Functions

Disassemble Entity
For each MAC-hs PDU that is delivered to the disassembly entity, the UE shall: remove any padding bits if present; remove the MAC-hs header; deliver the MAC-d PDUs in the MAC-hs PDU to MAC-d.
MAC-hs Reset
If a reset of the MAC-hs entity is requested by upper layers, the UE shall: flush soft buffer for all configured HARQ processes; stop all active re-ordering release timer (T1) and set all timer T1 to their initial value; start TSN with value 0 for the next transmission on every configured HARQ process; initialize the variables RcvWindow_UpperEdge and next_expected_TSN to their initial values; disassemble all MAC-hs PDUs in the re-ordering buffer and deliver all MAC-d PDUs to the MAC-d entity; flush the re-ordering buffer. and then: indicate to all AM RLC entities mapped on HS-DSCH to generate a status report. MAC-hs reset is usually performed in case of the serving hs cell is changed from one NodeB to another when the HARQ soft information is lost. This could also be applied for cell changes also within the same NodeB if there are other implementation restrictions in the NodeB causing soft information to be lost at cell change.
4-60
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-61
Version 1 Rev 0
HS-PDSCH and DL TrCH Information

The table shows what are the values and their need for following IEs Activation Time, H-RNTI, Downlink HS-PDSCH Information and Added or Reconfigured DL TrCH Information. Note, that only the IEs which have not yet been covered in the previous chapters will be explained here.
Activation Time
If the UE receives an RRC message where the IE "Activation time" has a value other (e.g. CFN) than the MD value "Now", the UE shall: in the case of HS-DSCH, let the CCTrCH including the associated DCH be the "reference CCTrCH", if the frame boundary immediately before the frame with the CFN value indicated by the IE "Activation Time" is at the TTI boundary common to all the transport channels that are multiplexed onto the reference CCTrCH: select that frame boundary as the activation time T. else: select the next TTI boundary, which is common to all the transport channels that are multiplexed onto the reference CCTrCH, after the frame with the CFN value indicated by the IE "Activation Time", as the activation time T. For an HS-DSCH related reconfiguration caused by the received RRC message: select the HS-SCCH subframe boundary immediately before the first HS-SCCH subframe, which entirely falls within the 10 ms frame following T; start using at that HS-SCCH subframe boundary the new HS-DSCH configuration in the received RRC message, replacing any old HS-DSCH configuration. Note: In FDD an "HS-DSCH related reconfiguration" includes, in particular, reconfigurations that need to be time-aligned with the 2ms subframe of the HS-SCCH, HS-PDSCH and/or HS-DPCCH. For example, start and stop of HS-SCCH reception and serving HS-DSCH cell change. If the UE receives a RRC message where the IE "Activation time" has the default value "Now", the UE shall: choose an activation time T as soon as possible after the reception of the message, respecting the performance requirements in 3GTS 25.331 (13.5.2) at the activation time T: perform the actions for the information elements in the received message as specified elsewhere
Downlink Transport Channel Type

The IE "Downlink transport channel type" values "HS-DSCH" and "DCH + HS-DSCH" are not used in the RRC CONNECTION SETUP message.
4-62
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-63
Version 1 Rev 0
HS-DSCH Data Frame on Iub/Iur (HS-DSCH FP)

The Data Transfer procedure is used to transfer a HS-DSCH DATA FRAME from the CRNC to a NodeB. When the CRNC has been granted capacity by the Node B via the HS-DSCH CAPACITY ALLOCATION Control Frame or via the HS-DSCH initial capacity allocation and the CRNC has data waiting to be sent, then the HS-DSCH DATA FRAME is used to transfer the data. If the CRNC has been granted capacity by the Node B via the HS-DSCH initial capacity allocation, this capacity is valid for only the first HS-DSCH DATA FRAME transmission. When data is waiting to be transferred, and a CAPACITY ALLOCATION is received, a DATA FRAME will be transmitted immediately according to allocation received. Multiple MAC-d PDUs of same length and same priority level (CmCH-PI) may be transmitted in one MAC-d flow in the same HS-DSCH DATA FRAME. The HS-DSCH DATA FRAME includes a User Buffer Size IE to indicate the amount of data pending for the respective MAC-d flow for the indicated priority level. Within one priority level and size the MAC-d PDUs shall be transmitted by the Node B on the Uu interface in the same order as they were received from the CRNC.
CmCH-PI (Common Transport Channel Priority Indicator)

CmCH-PI IE indicates the priority of the data frame and the SDUs included which are waiting in the SRNCs TX buffer for transmission via HS-DSCH, configured via the scheduling priority indicator in NBAP. Value range: {0-15, where 0=lowest priority, 15=highest priority}; Field length: 4 bits.
MAC-d PDU Length

Indicates the length of every MAC-d PDU in the payload of the HS-DSCH DATA FRAME in number of bits. Value range: {0-65535}; Field length: 16 bits.
NumOfPDU
Indicates the number of MAC-d PDUs in the payload. Value range: {1-255}; Field Length: 8 bits.
User Buffer Size

Indicates the users buffer size (i.e. the amount of data in the buffer) in octets for a given Common Transport Channel Priority Indicator level. Value range: {0-65535}; Field length: 16 bits.
MAC-d PDU
According to NumOfPDU the amount of MAC-d PDU is contained in the payload of HS-DSCH DATA FRAME.
4-64
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-65
Version 1 Rev 0
Control Frame Format for HS-DSCH FP

The flow control mechanism for HSDPA is based on two Control Frames: HS-DSCH Capacity Request; and HS-DSCH Capacity Allocation
HS-DSCH Capacity Request Control Frame

The HS-DSCH Capacity Request procedure provides means for the CRNC to request HS-DSCH capacity by indicating the user buffer size in the CRNC for a given priority level. The CRNC is allowed to re-issue the HS-DSCH Capacity Request if no CAPACITY ALLOCATION has been received within an appropriate time threshold. HS-DSCH Capacity Request is sent for each priority group to indicate the user buffer size. The control frame is sent for each priority group to indicate the user buffer size in SRNC (buffer occupancy). The control frame is sent by the SRNC when the SRNC considers the user buffer status needs an increased buffer reporting frequency. This may be sent to signal an event, such as, data arrival or user-buffer discard. This control frame is used to improve user-buffer reporting above the level produced by the user-buffer reporting associated with the HS-DSCH DATA FRAMEs.
4-66
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-67
Version 1 Rev 0
HS-DSCH Capacity Allocation Control Frame

HS-DSCH Capacity Allocation procedure is generated within the Node B. It may be generated either in response to a HS-DSCH Capacity Request or at any other time. The Node B may use this message to modify the capacity at any time, irrespective of the reported user buffer status. The CAPACITY ALLOCATION Control Frame describes an allocation that the SRNC may use. When the HS-DSCH Credits IE has a value of 0 it signifies that there is no resources allocated for transmission and to thus stop transmission. When the HS-DSCH Credits IE has a value of 2047, it signifies unlimited capacity for transmission of PDUs. When the HS-DSCH Repetition Period IE has a value of 0, it signifies that the allocation (Maximum MAC-d PDU Length, HS-DSCH Credits and HS-DSCH Interval IEs) can be repeated without limit.
Maximum MAC-d PDU Length

The value indicates the maximum allowable PDU size among the MAC-d PDU sizes configured via NBAP [6]. Value range & Field length: Refer to clause Data Frame Format for HS-DSCH FP on Iub/Iur.
HS-DSCH Credits
The HS-DSCH Credits IE indicates the number of MAC-d PDUs that a CRNC may transmit during one HS-DSCH Interval granted in the HS-DSCH CAPACITY ALLOCATION Control Frame. Value range: {0-2047, where 0=stop transmission, 2047=unlimited}; Field length: 11 bits.
HS-DSCH Interval
The value of this field indicates the time interval during which the HS-DSCH Credits granted in the HS-DSCH CAPACITY ALLOCATION Control Frame may be used. The first interval starts immediately after reception of the HS-DSCH CAPACITY ALLOCATION Control Frame, subsequent intervals start immediately after the previous interval has elapsed. This value is only applied to the HS-DSCH transport channel. Value range: {0-2550 ms}. Value 0 shall be interpreted that none of the credits shall be used; Granularity: 10ms; Field Length: 8 bits.
HS-DSCH Repetition Period

Description: The value of this field indicates the number of subsequent intervals that the HS-DSCH Credits IE granted in the HS-DSCH CAPACITY ALLOCATION Control Frame may be used. These values represent an integer number of Intervals. This field is only applied to the HS-DSCH transport channel. Value range: {0-255, where 0= unlimited repetition period}; Field Length: 8 bits.
4-68
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
4-69
Version 1 Rev 0
4-70
MOTOROLA LTD.2002
HSDPA Mobility
Version 1 Rev 0
Chapter 5 HSDPA Mobility
MOTOROLA LTD.2002
5-1
Version 1 Rev 0
HSDPA Mobility
5-2
MOTOROLA LTD.2002
Objectives
Version 1 Rev 0
Objectives
After this Lecture the Student will be able to: Describe the serving HS-DSCH cell change State the various HSDPA mobility procedures Describe the measurement event 1D State intra and inter NodeB serving HS-DSCH cell change
MOTOROLA LTD.2002
5-3
Version 1 Rev 0
Introduction to HSDPA Mobility Procedures

While in CELL_DCH state, the UE may be allocated one or more HS-PDSCH(s), allowing it to receive data on the HS-DSCH(s). Mobile evaluated hard-handover and soft-handover mechanisms are provided by the RRC connection mobility in CELL_DCH state. The mobility procedures are affected by the fact that the HS-PDSCH allocation for a given UE belongs to only one of the radio links assigned to the UE, the serving HS-DSCH radio link. The cell associated with the serving HS-DSCH radio link is defined as the serving HS-DSCH cell. A change in the serving HS-DSCH cell is the transfer of the serving HS-DSCH radio link from the source to the target HS-DSCH cell. Note: In Release 5, only network controlled serving HS-DSCH cell changes shall be supported. The SRNC makes the decision of the HS-DSCH cell change and determines the target cell. The decision could be based on UE measurement reports and other information available in the network. The SRNC controlled HS-DSCH cell change is performed as an RRC layer signaling procedure and is based on the existing handover procedures in CELL_DCH state.
Serving HS-DSCH Cell Change

The following categories exist with respect to the dedicated physical channel configuration: Serving HS-DSCH cell change while keeping the dedicated physical channel configuration and the active set; Serving HS-DSCH cell change in combination with an establishment, release and/or reconfiguration of dedicated physical channels Serving HS-DSCH cell change in combination with active set update in soft handover. With respect to synchronization between UE and UTRAN as to when transmission and reception is stopped and re-started, two possibilities for a serving HS-DSCH cell change exist: Synchronized serving HS-DSCH cell change: Start and stop of HS-DSCH transmission and reception is performed at a certain time typically selected by the network; Unsynchronized serving HS-DSCH cell change: Start and stop of HS-DSCH transmission and reception is performed "as soon as possible" (stated by UE performance requirements) at either side. Note: For traffic classes like streaming, interactive or background, the synchronized cell change is advantages as it minimizes the packet loss and therefore the jitter.
5-4
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
5-5
Version 1 Rev 0
Best Serving HS-DSCH Cell Measurement

The figure shows the event 1D which is used by UE to report a change of the best serving cell of the active set. In order to support the UEs mobility, the UE needs to provide measurement reports about the quality (Ec/Io) and / or signal level (RSCP) of the configured neighbor cells. Note this measurements are obtained from the P-CPICH which is basically a beacon for the cell.
Change of best Cell

As known from legacy releases, it is the SRNC that determines the cells which should belong to the UEs active set for transmission of dedicated channels. The SRNC typically bases its decisions on requests received from UE that are triggered by measurement events (e.g. Ec/Io or RSCP) on the P-CPICH. Via RRC measurement control messages the UE is informed about the candidate set for neighbor measurements. For HSDPA a similar measurement event 1D can be defined which is called the measurement event for best serving HS-DSCH cell. This measurement basically reports the best serving HS-DSCH cell to the SRNC based on a measurement of the P-CPICH Ec/Io or P-CPICH_RSCP. It is possible to configure this measurement event so that all cells in the UEs monitoring set are taken into account, or to restrict the measurement event so that only the current cells in the UEs active set (for dedicated channels) are considered. The usage of a hysteresis margin avoids fast change of the serving HS-DSCH cell. Also the specification of a CIO (cell individual offset) to favor certain cells, i.e. for instance, to extend their HSDPA coverage area.
HS-DSCH Handover
As can be seen in the figure, the UE moves from cell 1 towards cell 2. At a point in time, the P-CPICH_Ec/Io of cell 2 is better compared to cell 2s P-CPICH_Ec/Io. After a safe guard time of TTT (time to trigger), the UE reports event 1D to SRNC indicating cell 2s scrambling code and P-CPICH_Ec/Io and / or P-CPICH_RSCP value in order to inform SRNC about a better cell. TTT prevents in conjunction with H(1d) too frequent measurement reports and consequently too rapid cell changes. For example, the cell 2 must exceed cell 1s P-CPICH_Ec/Io quality level by H(1d) for a minimum time of TTT. Here we assume that CIO is set to 0 and thus not delaying or speeding up event 1D measurement reports. Once the SRNC has received the event 1D measurement report it may execute the HS-DSCH cell change procedure. Thus the HS-DSCH cell change is network controlled. The SRNC specifies with the activation time when the UE has to leave the old cell and change to the new cell. As this cell change involves RRC, NBAP and ALCAP signaling procedures in conjunction with an activation time, there is a certain delay until the serving HS-DSCH cell change is finished.
5-6
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
5-7
Version 1 Rev 0
Intra NodeB Synchronized HS-DSCH Cell Change

The picture shows how the serving HS-DSCH cell is changed within the same NodeB. For simplicity reasons only one MAC-hs function is shown although every cell has its own MAC-hs entity. Once the SRNC decides to make an intra NodeB handover from a source HS-DSCH cell to a new target HS-DSCH cell within the same NodeB, the SRNC sends via NBAP a Synchronized Radio Link Reconfiguration Prepare message to the NodeB followed by a (RRC) Physical Channel Reconfiguration message to the UE. At a specified time index (activation time) where the handover from the source cell to the new target cell is carried out, the source cell stops transmitting to the UE. Thereafter the MAC-hs packet scheduler is allowed to control from now on transmission to the UE. Similarly the UE starts listening to the HS-SCCH (up to four HS-SCCH at maximum, depending on the new cells circumstances in terms of channelization code resources and handling controlled by CRNC) from the target cell which is the new serving HS-DSCH cell. This also implies, that the UE estimates the downlink channel quality of the target cell. It is typically recommended, that the MAC-hs in the target cell does not start transmitting to the UE until it has received the first CQI report that is measured from the target cell. Prior to the HS-DSCH handover from the source cell to the new target cell, there are likely to be several PDUs buffered in the source cells MAC-hs for the UE. This PDUs comprise of PDUs that have never been sent transmitted to the UE and pending PDUs in the HARQ manager that are either awaiting to be Ack/Nacked on the uplink HS-DPCCH or PDUs that are waiting to be retransmitted to the UE.
MAC-hs Preservation
MAC-hs preservation means that in case of intra NodeB HS-DSCH serving cell change all the PDUs for the UE are moved from the MAC-hs in the source cell to the MAC-hs in the target cell. This means that the status of the HARQ manager is also preserved without triggering any higher layer retransmission such as RLC retransmission during intra NodeB HS-DSCH to HS-DSCH handover. If the NodeB does not support MAC-hs preservation, then the handling of the not completed PDUs is the same as in the inter NodeB case. Note: In the case of RLC-AM, the polling function can be used to obtain the status of the data transmission to the UE at the RLC level. So the SRNC can immediately forward all not yet acknowledged RLC PDUs to the new cell that were deleted in the source cell prior to the HS-DSCH cell change. In the case of RLC-UM the PDUs in the old cell waiting for transmission might be transferred to the target cell in case of intra NodeB HS-DSCH cell change but this depends on vendor implementation. During intra NodeB HS-DSCH to HS-DSCH handover, it is likely, that the UEs DPCH is in soft handover. It could be possible that the HS-DPCCH is also in soft handover condition, so both cells of the NodeB allocate Rake fingers for demodulation and receive a combined HS-DPCCH signal. This could improve uplink coverage of the HS-DPCCH and no power control problems are expected compared to the Inter NodeB soft handover.
5-8
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
5-9
Version 1 Rev 0
Intra NodeB Synchronized HS-DSCH Cell Change Procedure

The figure illustrates an intra-Node B serving HS-DSCH cell change while keeping the dedicated physical channel configuration and the active set, using the Physical Channel Reconfiguration procedure. The transition from source to target HS-DSCH cell is performed synchronized, i.e. at a given activation time. In this example, the UE transmits a Measurement Report message containing intra-frequency measurement results, here assumed to be triggered by the event 1D "change of best cell". When the SRNC has performed the handover decision, the Node B is prepared for the serving HS-DSCH cell change at an activation time indicated via a NBAP RL Reconfiguration Commit message. The SRNC then sends a Physical Channel Reconfiguration message, which indicates the target HS-DSCH cell and the activation time to the UE. Since the same Node B controls both the source and target HS-DSCH cells we assume there is no need to reset the MAC-hs entities. Note MAC-hs reset is usually performed when the HARQ soft information is lost (status about not yet Ack/Nacked MAC-hs PDUs). This could happen during cell changes also within the same NodeB. When the UE has completed the serving HS-DSCH cell change it transmits a Physical Reconfiguration Complete message to the network. In this example it is assumed that HS-DSCH transport channel and radio bearer parameters do not change. If transport channel or radio bearer parameters shall be changed, the serving HS-DSCH cell change would need to be executed by a Transport Channel Reconfiguration procedure or a Radio Bearer Reconfiguration procedure, respectively. 1. 2. UE sends Measurement Report [E1D] to SRNC via DCCH (on DPCH) in order to indicate a better cell. Measurement reporting criteria could be e.g. P-CPICH Ec/Io. The SRNC decides there is a need for a serving HS-DSCH cell change and requests the serving HS-DSCH NodeB to perform a synchronized radio link reconfiguration using NBAP message Synchronized Radio Link Reconfiguration Prepare. Parameters provided: HS-DSCH Information, HS-DSCH RNTI and HS-PDSCH RL ID. The serving HS-DSCH NodeB returns a NBAP message Synchronized Radio Link Reconfiguration Ready. Parameters provided: HS-DSCH Information Response. The SRNC now proceeds by transmitting NBAP message Radio Link Reconfiguration Commit to NodeB. Parameters provided: SRNC selected activation time in the form of CFN. SRNC transmits a RRC message Physical Channel Reconfiguration to the UE. Parameters provided: Activation time, MAC-hs reset indicator, serving HS-DSCH radio link indicator, HS-SCCH set info and H-RNTI. At the indicated activation time the UE stops receiving HS-DSCH in the source cell and starts HS-DSCH reception in the target cell. The UE then returns a Physical Channel Reconfiguration Complete message to the SRNC via DCCH (on DPCH).
3. 4. 5.
6.
5-10
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
5-11
Version 1 Rev 0
Inter NodeB Synchronized Serving HS-DSCH Cell Change

The figure depicts Inter NodeB HS-DSCH to HS-DSCH handover. In this scenario the serving HS-DSCH NodeB is under one NodeB while the new target cell is under another NodeB. For simplicity reasons we show the Intra RNC case. The Inter NodeB HS-DSCH serving cell change could also take place between two RNCs (SRNC DRNC) if the target NodeB is under control of another RNC. Once the SRNC decides to initiate such a handover (e.g. after UE reported event 1D) a NBAP Synchronized Radio Link Prepare message is sent to the new NodeB that controls the target cell. Afterwards a RRC Transport Channel Reconfiguration is sent to the UE which indicates to the UE that the MAC-hs entity should be reset and a status report for each RLC-entity associated with the HS-DSCH should be generated. At the time, the cell change should be executed, the MAC-hs for the UE in the source cell is reset, which basically means that all buffered MAC-d PDUs for the UE are deleted including the pending MAC-hs PDUs in the HARQ manager. At the same time index (activation time), the flow control unit in the MAC-hs in the target cell starts to request PDUs from MAC-d of the SRNC, so that it can start to transmit data on the HS-DSCH to the UE. As the old MAC-d PDUs, that were buffered in the source cell prior to the handover, are deleted, these PDUs must be recovered by higher layer retransmission such as RLC retransmission. When the RLC protocol realizes, that the RLC PDUs it has originally forwarded to the source cell are not acknowledged, it will initiate retransmission, which results in forwarding the same PDUs via MAC-d to the new target cell. In order to reduce the potential PDU retransmission delays during recovery phase, the RLC protocol at the UE can be configured to send RLC status report to the SRNC at the first time incident after the serving HS-DSCH cell has been changed. For user applications that do not include any higher layer retransmission mechanisms, e.g. applications running over UDP and use RLC-UM, the PDUs that are deleted in the source cells MAC-hs prior to the handover are lost. For applications having large data (MAC-d PDUs) amounts buffered in MAC-hs should therefore be avoided, as these may be lost in an inter NodeB serving cell change is suddenly initiated. The consequence out of this should be that the flow control between MAC-hs and MAC-d gets optimized. The flow control should ensure that the MAC-hs buffer has just sufficient MAC-d PDUs (RLC-UM PDUs) to allow for continues data transmission on HS-DSCH. Such a optimized flow control avoiding the deletion of large MAC-d PDUs in case of inter NodeB HS-DSCH cell change impacts also the scheduler in MAC-hs. So there is a lot of room for proprietary improvements.
5-12
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
5-13
Version 1 Rev 0
Inter NodeB HS-DSCH Cell Change Hard Handover(1)

The figure illustrates a synchronized inter-Node B serving HS-DSCH cell change in combination with hard handover e.g. necessary for quick handovers of both DCHs and HS-DSCH in one shot. The reconfiguration is performed in two steps within UTRAN. On the radio interface only a single RRC procedure is used. The UE sends Measurement Report with event 1D to the SRNC. The SRNC determines the need for hard handover based on received measurement reports and/or load control algorithms ( e.g. measurements may be performed in compressed mode Inter FDD HHO). In the first step, the SRNC establishes a new radio link in the target NodeB via a NBAP RL Setup message. In the second step this newly created radio link is prepared for a synchronized reconfiguration to be executed at a given activation time indicated in a NBAP RL Reconfiguration Commit message. After the first step, the target Node B starts transmission and reception on dedicated (logical) channels (DCH DCCH). At the indicated activation time, transmission of HS-DSCH is started in the target HS-DSCH Node B and stopped in the source HS-DSCH Node B. In order to trigger the hard handover SRNC then sends a Transport Channel Reconfiguration message on the old configuration. This message indicates the configuration after handover, both for DCH and HS-DSCH. The Transport Channel Reconfiguration message includes a flag indicating that the MAC-hs entity in the UE shall be reset. The message also includes an update of transport channel related parameters for the HS-DSCH in the target HS-DSCH cell. 1. 2. 3. UE sends Measurement Report [E1D] to SRNC via DCCH in order to indicate a better cell. The SRNC decides that there is a need for a hard handover combined with a serving HS-DSCH cell change. It therefore transmit a NBAP Radio Link Setup message to the target NodeB. The target NodeB allocates the resources, starts physical layer reception on the DPCH on the new radio link and responds with the NBAP message Radio Link Setup Response. Parameters provided: HS-DSCH Information Response. The SRNC initiates set-up of a new Iub Data Transport Bearer for DCH using ALCAP protocol. This request contains the AAL2 Binding ID to bind the Iub Data Transport Bearer to the DPCH. The SRNC requests the source HS-DSCH NodeB to perform a Synchronized Radio link Reconfiguration, removing its HS-DSCH resources for the source HS-DSCH radio link. Parameters provided: HS-DSCH Information, a CRNC (=SRNC) allocated HS-DSCH RNTI and HS-PDSCHs RL ID. The source HS-DSCH NodeB returns the NBAP message Synchronized Radio Link Reconfiguration Ready. Parameters provided: HS-DSCH Information Response. The SRNC transmits a NBAP Radio Link Reconfiguration Commit to the source cell indicating when the MAC-hs shall stop sending HS-DSCH data blocks. At the indicated activation time the source NodeB stops and the target HS-DSCH NodeB starts transmitting on the HS-DSCH to the UE. Parameter provided: SRNC selected activation time in the form of CFN. Now the SRNC requests the target HS-DSCH NodeB to perform a Synchronized Radio Link Reconfiguration via NBAP, adding HS-DSCH resources for the target HS-DSCH radio link. The target HS-DSCH NodeB returns the NBAP message Radio Link Reconfiguration Ready. Parameters provided: HS-DSCH Information Response.
4. 5.
6. 7.
8. 9.
5-14
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
5-15
Version 1 Rev 0
Inter NodeB HS-DSCH Cell Change Hard Handover (2)

The UE terminates transmission and reception on the old radio link at the activation time indicated in the Transport Channel Reconfiguration message, and configures its physical layer to begin reception on the new radio link. After L1 synchronization in the new cell has been established, the UE sends a Transport Channel Reconfiguration Complete message. The SRNC then terminates reception and transmission on the old radio link for dedicated channels (DCH DCCH/DTCH) and releases all resources allocated to the considered UE. Note in this inter-Node B handover example, RLC-PDUs transmission/reception via HS-DSCH is stopped at both the UTRAN and UE sides prior to reconfiguration and continued when the reconfiguration is completed. It is furthermore assumed in this example that the Transport Channel Reconfiguration message indicates to the UE that the MAC-hs entity should be reset and a status report for each RLC entity associated with the HS-DSCH should be generated. A reset of the UE MAC-hs entity triggers the delivery of the content in the re-ordering buffer to higher layers. As the UE has changed the NodeB, the retransmission of faulty MAC-hs PDUs being Nackd in the source cell are not possible in the new cell. Therefore soft combining of faulty received MAC-hs PDUs is not possible anymore as the new cell does not get the HARQ preservation info like it is possible in the Intra NodeB cell change case. 10. The SRNC initiates the set-up of a new Iub Data Transport Bearer using ALCAP protocol. This request contains AAL2 Binding ID to bind the Iub Data Transport Bearer to the HS-DSCH. 11. The HS-DSCH transport bearer to the target HS-DSCH NodeB is established. The SRNC proceeds by transmitting the NBAP Radio Link Reconfiguration Commit to NodeB. Parameters provided: SRNC selected activation time in the form of a CFN. 12. The SRNC transmits a RRC message Transport Channel Reconfiguration to the UE. Parameters provided: Activation time, MAC-hs reset indicator, serving HS-DSCH radio link indicator, HS-SCCH set info and H-RNTI. 13. At the indicated activation time the UE abandons the current active set and initiates establishment of the DPCH in the target cell.. When physical layer synchronization is established in the target cell, UE starts DPCH reception and transmission and HS-DSCH reception in the target cell. The UE returns a RRC message Transport Channel Reconfiguration Complete to the SRNC. 14. The SRNC finalizes the procedure by transmitting a Radio Link Deletion Request to the source NodeB to de-allocate the radio resources (DPCH and HS-PDSCHs). 15. The source NodeB releases the HS-DSCH and DCH resources and returns a Radio Link Deletion Response to the SRNC. Finally the DCH and HS-DSCH Transport Bearers are released through ALCAP.
5-16
MOTOROLA LTD.2002
Version 1 Rev 0
MOTOROLA LTD.2002
5-17

Sys205!1!0 (Hsdpa Protocols)

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Sys205!1!0 (Hsdpa Protocols)

Uploaded by

Copyright:

Available Formats

HSDPA System Architecture and Protocols

HSDPA System Architecture and Protocols Chapter 1: HSDPA Overview

HSDPA System Architecture and Protocols

HSDPA System Architecture and Protocols

Chapter 2: HSDPA Principles

HSDPA System Architecture and Protocols

HSDPA System Architecture and Protocols

HSDPA System Architecture and Protocols

2-100 2-100 2-100 2-100

Chapter 3: Forward and Backward Error Correction

HSDPA System Architecture and Protocols

Chapter 4: HSDPA Protocol Enhancements and Extensions

HSDPA System Architecture and Protocols

HSDPA System Architecture and Protocols

Chapter 5: HSDPA Mobility

HSDPA System Architecture and Protocols

This page intentionally left blank.

About This Manual

HSDPA System Architecture and Protocols

This page intentionally left blank.

Chapter 1 HSDPA Overview

This page intentionally left blank.

HSDPA (High Speed Downlink Packet Access) for WCDMA

HSDPA (High Speed Downlink Packet Access) for WCDMA

Compatibility with Release 99

Demand for Packet Switched Traffic

HSDPA (High Speed Downlink Packet Access) for WCDMA

HSDPA (High Speed Downlink Packet Access) for WCDMA

Consideration of UE Processing Time and Memory Requirements

Higher Spectrum Efficiency

Small Changes to existing Techniques and Architectures

Efficient Resource Sharing in Downlink among Users

Higher Throughput Rates

AMC (Adaptive Modulation and Coding)

Transmission and Retransmission Scheduling in NodeB

QPSK versus 16-QAM Modulation

QPSK versus 16-QAM Modulation

QPSK versus 16-QAM Modulation

QPSK versus 16-QAM Modulation

Maximum Throughput Rates

Maximum Throughput Rates

UMTS Rel. 99 / Rel. 4

Maximum Throughput Rates

Maximum Throughput Rates

UMTS Rel. 99 / Rel 4

Important Changes for HSDPA

Important Changes for HSDPA

New 2 ms Subframe for HSDPA

New Physical Channels and Transport Channel with HSDPA

No Fast Power Control and variable Spreading Factor

New UE Capabilities / Categories

New MAC-hs in NodeB and UE

Impact on NBAP and Frame Protocol Procedure

Important Changes for HSDPA

Important Changes for HSDPA

New Channels with HSDPA

New Channels with HSDPA

New Channels with HSDPA

New Channels with HSDPA

HSDPA Basic Operation

HSDPA Basic Operation

HSDPA Basic Operation