ALCATEL ConnectMe Broadband

CONNECT ME ANYWHERE, ANY TIME, ANY WAY ...
ALCATEL TELECOMMUNICATIONS REVIEW
Alcatel Telecommunications Review - 2 nd Quarter 2005
UNIVERSAL ACCESS
CONTENTS
Alcatel Telecommunications Review is the quarterly technical journal of Alcatel, reporting its research, development and production achievements worldwide.
Connect me anywhere, any time, any way
EDITORIAL
Project Lightspeed Ernie Carey
90
INTRODUCTION
M. Peruyero, Y. TJoens
EDITORIAL BOARD
Niel Ransom Alcatel, Paris, France Ron Spithill Alcatel, Paris, France Olivier Baujard Alcatel, Paris, France Jolle Gauthier Alcatel, Paris, France Pierre Tournassoud Alcatel, Paris, France Alistair Urie Alcatel, Paris, France Vince Pizzica Alcatel Asia Pacific, Shanghai, Peoples Republic of China Guido H. Petit Alcatel, Antwerp, Belgium
92
Universal Broadband Access

Making User-Centric Broadband in Access a Reality
A seamless combination of fixed, mobile and broadcast access offers users an always best connected experience.
94
Ph. Lain, L. Le Gouriellec, J. De Vriendt
Any Access
Universal Broadband Access: Going Wireless and Mobile 99
Are the emerging broadband wireless and mobile technologies a threat or an opportunity? Will they compete or complement one another?
J-L.Hurel, J. Brouet, L. Le Gouriellec, M. Peruyero Software Defined Radio: A Promising Technology for Multi-Standard Base Stations 106
GUEST EDITOR
Pierre Tournassoud Alcatel, Paris, France
EDITORS
Willem Zevenbergen Editor in Chief, Paris, France Catherine Camus Managing Editor, Paris, France
SDR is an important step towards offering every user at any location the best bandwidth and mobility, in line with Alcatels vision of a usercentric broadband world.
B. Haberland, W. Koenig, A. Pascht, U. Weiss Standardization: Key to Mass Deployment

Standards have evolved to make DSL technology a credible building block for triple-play deployments.
112
CONTRIBUTING EDITORS
Gustavo Arroyo Spanish edition, Madrid, Spain Mike Deason English edition, Braintree, UK
F. van der Putten, S. Ooghe Optical Fibers Pave the Way to Faster Broadband Access
Emerging services and increasing competition are forcing carriers to deploy optical fibers in the access network, ultimately serving each subscriber over a dedicated fiber link.
116
Th. Pfeiffer, E. Ringoot, A. Granger, D. Wang
Articles marked @ are the web supplement to this edition at www.alcatel.com/atr

www.alcatel.com/atr
Extending Broadband Reach by Satellite

Combining broadband satellite solutions with terrestrial wireline and wireless solutions can rapidly extend the reach of IP multimedia applications to underserved parts of the world.
122
Alcatel Evolium Multi-standard Radio Access System

Multi-standard radio access is the new trend in network design! So what are the requirements and benefits?
150
F. Engmann, G. Walz
A. Bertout, J. Couet
Operational Excellence Any Service

Supporting Quality of Service in Broadband Access Networks 128 Optimizing DSL for Multimedia Services
Line optimization can increase DSL speed, quality and stability while reducing operating expenses.
155
An optimal user experience for triple-play services can be assured thanks to QoS enablers in fixed and wireless access networks.
T. Bostoen, R. Oehen, J. Verlinden 4G Mobile

4G will deliver low cost multi-megabit/s sessions any time, any place, using any terminal.
S. Ooghe, N. Drevon, R. Siebelink Access Network Enhancements for the Delivery of Video Services 134
160
D. Rouffet, S. Kerboeuf, L. Cai, V. Capdevielle
Enhancements are being researched that will improve both the quality of video delivered over networks experiencing congestion or transmission errors, and the video channel changing performance.
R. Sharpe, D. Zriny, D. De Vleeschauwer
Customer Example
3G Network Powers Intense Services 165
The launch of Orange Intense marks the culmination of three successful years of close collaboration with Alcatel.
Consolidation
Fixed Access Vision 140
Guaranteed delivery of triple-play services making optimal use of valuable resources is the key driver moving fixed access forward as the world converges towards offering all services over ubiquitous broadband connectivity.
O. Perot, J-M. Perera, L. Byerley
J. van Bogaert, Y. TJoens, J-P. Lartigue WiMAX: From Fixed Wireless Access to Internet in the Pocket
WiMAX access can provide new service opportunities to fixed and mobile operators thanks to its flexible radio technology and innovative features.
144
D. Renaudeau, D. Boettle, H. Steyaert
Customer Applications Notes www.alcatel.com/atr
Technology White Paper
Strategy White Paper
Technical Paper
2 nd Quarter 2005 - Alcatel Telecommunications Review -
89
EDITORIAL
Ernie Carey
SBC PROJECT LIGHTSPEED
Ernie J. Carey Vice President, Network, IP Operations & Services SBC Inc
BC companies are bringing a digital lifestyle to customers. As customers expect more bundled services, enriched video, data and voice experiences, SBC is moving quickly to offer next-generation communications and entertainment services. The success of these services depends on networks working together to help SBC companies compete in the Internet Protocol (IP) marketplace. Widespread availability of broadband is a key element in making the digital lifestyle a reality. Consumer broadband adoption levels are reaching critical mass. In fact, according to Yankee Group, at the end of 2004, about 30 million US households had broadband service. At the end of 2004, SBC companies reached a significant milestone of surpassing 5 million DSL lines in service. Increasing the bandwidth of broadband is critical to meeting the rapidly evolving demands of the marketplace that is, to provide IP-based services. Why is now the right time? Today, traditional telephone companies like SBC are competing against a variety of providers and technologies by offering bundles that include voice, video, data and wireless. Meanwhile, competitors are starting to enter traditional SBC markets. For example, cable companies, which have offered video and high-speed Internet service for years, are now getting into the voice business. Some are offering Voice over Internet Protocol (VoIP) a technology thats receiving a lot of buzz. Theres no doubt that the marketplace is growing increasingly competitive. At the same time, customer needs are changing. Today were seeing an appetite for cooler, smaller, faster and better digital devices. Whether its a wireless phone that can shoot video, or a PDA that doubles as a portable music player, or a PC that can be used as a media center and digital video recorder consumers want maximum functionality and flexibility, and they want many of these devices integrated so they work simply together. The digital lifestyle is no longer just a buzzword. Its a reality. So, by fully utilizing the capabilities of IP, fiber and wireless data technologies, SBC companies are working to build an even stronger presence in the marketplace with differentiated, competitively priced product offerings. IP services will enable integrated video, voice, data and other applications, potentially allowing customers to share any number of household devices such as TVs, set-top boxes, PCs, PDAs or phones over one network connection.
SBC companies expect to add approximately 38 000 miles of fiber through Project Lightspeed, an estimated $4 billion capital initiative to deploy both Fiber to the Node (FTTN) and Fiber to the Premises (FTTP) to 18 million homes by the end of 2007. In most new network builds, such as developing subdivisions, SBC companies are planning to bring fiber all the way to the customers homes. In neighborhoods with existing SBC services, SBC companies plan to use FTTN technology to push fiber much deeper in the network to nodes that serve 300 to 500 homes. FTTN enables significantly higher broadband speeds, with download speeds of 20 to 25 Mbit/s. Cost, demand, time for deployment and avoiding potential inconvenience for customers are all key factors in this decision to use FTTN for overbuilds. SBC companies have evaluated a full range of technologies and deployment scenarios and are confident that the joint FTTN/FTTP strategy is the right solution to deliver the next generation of IP services, and to evolve the SBC network to meet customers communications needs. To help build this new network, SBC announced a five-year, approximately $1.7 billion primary supplier agreement with Alcatel in October 2004 to provide network equipment and video system integration services. Alcatels network equipment will include core network access, aggregation and switching equipment platforms that will provide the IP, packet-based technologies over fiber optics that connect the neighborhoods to the central office. Additionally, Alcatel will work with SBC to ensure video systems integration. For the video network, instead of using a traditional legacy broadcast network which requires all television content to be transmitted to every customers set-top box all the time, SBC companies are conducting trials of a switched video distribution system. In the switched video environment, the content resides on SBC servers and only the content the customer requests is sent, freeing up bandwidth to be used for other applications. SBC announced a contract with Scientific-Atlanta to provide IP-based video equipment for Project Lightspeed. Alcatel will work with SBC and now Scientific-Atlanta to ensure seamless video systems integration. Additionally, last year, SBC announced a first-of-its-kind agreement with Microsoft Corp to use the Microsoft TV Internet Protocol Television (IPTV) Edition software platform. SBC Labs has been testing an IP-based television service built on the Microsoft TV IPTV Edition platform since June 2004.
90
- Alcatel Telecommunications Review - 2 nd Quarter 2005
www.alcatel.com/atr
SBC PROJECT LIGHTSPEED
SBC companies expect that this video distribution system coupled with its fiber-rich network will allow it to offer an IPTV service that offers features and functionality beyond todays existing broadcast-delivered digital TV networks. For example, in the future the SBC service could include a robust video-on-demand library, multiple picture-in-picture functionality, whole-house digital video recording, interactive program guides and more. And, with the remaining bandwidth, SBC plans to offer feature-rich VoIP and high speed Internet access. Some additional potential applications include: Using two-way broadband networks, SBC companies could be able to send alerts and notifications to customers watching TV in new ways. Some examples include the ability to alert a customer of upcoming favorite shows, or Caller ID and instant messaging on the TV screen.
The Microsoft TV IPTV Edition platform could enable new services and applications, such as tunerless picture-in-picture functionality. The picture-in-picture feature enables users to preview other shows and channels while the primary channel runs in the background. IP is the future, and SBC companies are embracing the future to deliver next-generation communications and entertainment services.
This article was written as of April 2005, and SBC disclaims any obligation to update or revise statements contained in this article based on new information or otherwise.
Ernie Carey began his career with Southwestern Bell in 1975, in Houston Texas, after graduating from college. He then progressed through a series of operations, engineering, and marketing jobs in Southwestern Bell, and is currently Vice President-Network with responsibility for network planning and engineering for Project Lightspeed. Mr Carey is on the Board of Directors of the Sam Houston Council of the Boy Scouts of America. He was appointed as a member of the Commission on State Emergency Communications and also served on the Board of Directors of the Greater Harris County E911 District for eight years. Currently he is on the board of the Texas Technology Opportunity Institute and the Engineering Advisory Board, Cullen College of Engineering, University of Houston.
www.alcatel.com/atr
91
INTRODUCTION
Michel Peruyero, Yves TJoens
INTRODUCTION
o realize a user-centric broadband world, service providers across the world will need to change their organizations and business models and to upgrade their networks. This transformation is being made necessary by the need of todays users to communicate more, while being less occupied with the underlying technologies. The initial step towards a user-centric broadband world is a ubiquitous broadband connection over which content and services can be made universally available. Connectivity is one of the users most basic requirements. This goes beyond broadband everywhere to a demand that access networks be invisible to the user, who doesnt want to have to choose which device, service and access to use. This implies the disappearance, in the users eyes, of distinctions between fixed, mobile and wireless access technologies. While this does not necessarily mean network convergence, it does entail service convergence from the users point of view (i.e. same device, same content, one bill, regardless of the access network), as well as seamless handovers between access networks. It also implies that broadband access technologies should become increasingly complementary. Broadband wireless technologies, such as WiMAX, will complement both mobile and fixed access, giving mobile users higher bandwidth data delivery in hotzones, and giving nomadic users DSL-equivalent service in rural and suburban areas. The key to a seamless broadband experience will be smooth handovers. Network operators facing the build-up of this new network are increasingly looking to reduce their overall operating costs. As well as investing in tools to ensure operational excellence, one will witness the first signs of network consolidation around this new paradigm of communication, with operators seeking to switch off their existing legacy telephony networks to consolidate all operations around a multimedia IP network, thus sowing the seeds for full fixed/mobile convergence. In this issue of the Alcatel Telecommunications Review, the emphasis is on Universal Broadband Access (see Figure 1). The article Making User-Centric Broadband in Access a Reality introduces the vision of future access network evolution with respect to fixed, mobile and broadcast access. The evolution of networks in general, and access networks in particular, in line with the universal access vision involves four dimensions: Any Access, Any Service, Consolidation and Operational Excellency. The Any Access dimension is paving the way to ubiquitous service delivery, which implies the availability of access
92
technologies that are fit for the purpose (and seamless roaming between different access technologies), as well as multiservice support using QoS techniques and intelligent network support for data, voice and video services. Access technologies are proliferating, giving operators a diversity of choice to fit their architectures and service delivery models. This issue provides a broad overview of Alcatels access technologies, which illustrate the innovation power of Alcatel. These include Mobile and wireless: Universal Broadband Access: Going Wireless and Mobile and Software Defined Radio: A Promising Technology for Multi-standard Base Stations. DSL: Optimizing DSL for Multimedia Services and Standardization: Key to Mass Deployment. Fiber: Optical Fibers Pave the Way to Faster Broadband Access. Satellite: Extending Broadband Reach by Satellite. The second dimension is Any Services. Multi-service support across fixed and mobile networks is detailed in Supporting Quality of Service in Broadband Access Networks, which focuses on how different service streams receive the treatment they need across the access network. One particular service that benefits from intelligent access network enhancements is the video service; intelligent traffic
Figure 1: Universal broadband access
Every Terminal, Everywhere
Business Environment
Universal Broadband Access Wireline
Consumer Environment
Wireless Satellite Mobile
Service Aware Edge & Data Aware Transport
Open Services Delivery Environment
Operations & Business Support Systems Integration & Operations
www.alcatel.com/atr
Any Content
Each User
management and advanced channel changing techniques enable operators to provide their customers with a unique experience. Recent innovations following on from Alcatel research are highlighted in Access Network Enhancements for the Delivery of Video Services. The third dimension of the universal access vision is based on Consolidation. Network consolidation is happening at various levels of the network, witnessed by the desire of operators to evolve their networks under themes such as fixed/mobile convergence and voice/data convergence. Voice/data convergence (Fixed Access Vision) using the newest generation of broadband loop carriers is fundamental to Alcatels view of how it can help operators to extend their offered services while controlling their capital and operational expenditures by means of network consolidation. The article WiMAX: From Fixed Wireless Access to Internet in the Pocket explains how WiMAX can be used as an enabler for fixed/mobile convergence. Furthermore, support for multiple radio technologies on a common infrastructure results in the consolidation of network equipment, leading to more costeffective solutions, as explained in Alcatel Evolium Multistandard Radio Access System. The fourth dimension of the universal access vision turns around Operational Excellence. Scaling the access network
in line with the proliferation in access technologies, the number of attached subscribers and the multiplication of service bundles, requires significant investments in service support systems. The tools offered in this respect for copper access networks are discussed in Optimizing DSL for Multimedia Services. Operational excellence is also a key driver for the future evolution of radio networks (4G Mobile). The aim is to realize service and application ubiquity, with a high degree of personalization and synchronization between the various appliances. This issue concludes with a real life network example which demonstrates how Alcatel is at the forefront in realizing the Universal Access Vision. This example covers the launch of Orange Intense (3G Network Powers Intense Services); it traces the Alcatel and Orange partnership from the pre-commercial phase to the commercial launch of Oranges 3G services offering. We hope that youll enjoy this issue, and that it will give you a valuable overview of the User-Centric Broadband world, with a special focus on the access part. Be ready for the rapidly approaching era of universal broadband access!
Michel Peruyero, Yves TJoens
Michel Peruyero is Executive Director in the Chief Technology Office of the Alcatel Mobile Communications Group, Vlizy, France. (Michel.peruyero@alcatel.fr) Yves TJoens is Chief Technology Officer for the Alcatel Access Network Division. Prior to obtaining this position he acted as corporate strategy director for access, and ran the research group on access networking architectures. He is the author of multiple papers and actively contributed to the standardization of access networking in the areas of IP/Ethernet and the ATM access architecture. He is based in Antwerp, Belgium.
www.alcatel.com/atr
93
STRATEGY WHITE PAPER
UNIVERSAL BROADBAND ACCESS
Ph. Lain, L. Le Gouriellec, J. De Vriendt
MAKING USER-CENTRIC BROADBAND IN ACCESS A REALITY

A seamless combination of fixed, mobile and broadcast access offers users an always best connected experience.
Figure 1: Applications should travel with users wherever they are and whatever they do n entire range of new services emerged (e.g. mobile voice, Internet access) from a communication world that used to be Anywhere defined by simple fixed voice services. Else: Mobile Then the emergence of broadband enabled users to access all these new services more rapidly. Nevertheless, for Hot Zone users the promise of ubiquitous broadNomadic (WiFi, WDSL) band is far from being fulfilled. The proliferation of terminals (fixed phone, mobile phone, PC, personal digital Home assistant, etc) has been further comFixed (DSL, Eth., PON, plicated by the abundance of broadWDSL) band access options, including Digital Subscriber Line (DSL), cable, UniverMy My Practical Life My Clan/Tribe Office sal Mobile Telecommunications System Entertainment Toolset with me Fixed with me with me Convenience, Productivity (UMTS), Code Division Multiple (DSL, Eth., PON, WDSL) Access 2000 (CDMA2000), public My Profile with me Wireless Local Area Networks (WLAN) and satellite. These multiple access My Office technologies generate overlapping with me functionality with multiple accounts, My Profile & multiple subscriptions, multiple bills IM: Instant Messaging PON: Passive Optical Network VPN with me and multiple user names and passPTT: Push To Talk WDSL: Wireless DSL words. So, while broadband has opened up a new world with numerous appealing services, the delivery of these new and richer services Consumer Demands has come at the cost of greater The user-centric broadband experience is about the complexity and a fragmented user being able to use all the services to which he or experience linked to the various she has subscribed from any location, using any access technologies. device, even when moving from a location served by one To improve the users expeaccess technology to a location served by a different access rience, there is a move towards technology (see Figure 1). simplicity and convenience. The network should select the access solution that is most Here we review users needs convenient for the user and minimizes his or her costs. and describe some access soluTo fully meet the users expectations, these principles should tions that operators can deploy apply to all the applications he or she needs during a normal to make users lives easier by days activities. enabling them to access their Users should be able to access their services from anywhere applications via the best availat any time. Moreover, while running a given application, a user able network at the highest should be able to move from one place to another without possible data rate, all through experiencing any disruption. To achieve this, the new access a single subscription. network should support a minimum set of features and a minVoice & video Call/conference/IM/PTT Music/Video/TV Multicast/Broadcast VPN voice & video Call/IM/PTT/Net-meeting VPN Data: Email/SMS/MMS, Intra/Internet Email/SMS/MMS, Internet, Purchase Email/SMS/MMS Voice Calls/Chat Gaming Music
94
www.alcatel.com/atr
Figure 2: Five services classes and their requirements
AP Hot Zone
AC Corporate Voice & Data VPN Multi/Unicast/Broadcast Video/TV/Music Combinational e.g. Multiplayer Gaming Calls & Conferencing Voice & Video & IM & PTT Inter/Intranet+email/MMS Leased Line QoS, Security Predictability, Availability Strict Multicast QoS Broadcast Capacity Interactivity QoS High Capacity Near or Real Time, Low Delay High Availability Best Effort No Impact
One or More Access Networks
Mobile
Fixed: DSL, Cable, etc.

AC: Access Controller AP: Access Point
Access Network (3G, WiFi/WiMAX, Fixed DSL, Cable) Application
Requirements To Access Networks
imum capacity. Consequently, it is important to analyze the mandatory requirements that different applications place on an access network. Main types of service All applications ultimately fall into five service classes, as shown in Figure 2. Data services: Internet access and messaging These applications deal with Internet access, file downloading, and online usage of applications or services. They include all call services operated by Internet Service Providers (ISP), such as Skype, that use the Internet in a transparent way, as well as access to email, the Short Message Service (SMS) and the Multimedia Messaging Service (MMS). Conversational services: Carrier-grade voice/video calls and conferences This class of applications addresses peoples need to communicate in real-time or near real-time with one or more others, using voice, video or messaging. The key to such applications is the ability to provide continuous reachability over a variety of access technologies while the user is on the move. These services can be classified as follows:
Real-time voice and video calls and conferencing, whether based on circuit switching or Voice over Internet Protocol (VoIP). Fast handover is crucial to ensure a good user experience. An acceptable round-trip delay is also important in order to offer carrier-grade quality in all circumstances. Near-real-time rich call services, such as instant messaging, push to talk, push to view, see-what-I-see and conferencing tools. Video services: TV, video and music streaming Applications of this type address the need of users to access video services, such as TV and video on demand, and multimedia content, in most cases using a streaming mode. The access network must meet several criteria to support this type of service:
Push To Talk (PTT) A walkie-talkie like service that provides voice messaging to a closed user group at the touch of a button. PTT uses voice-over-packet techniques to minimize spectrum usage. PTT is one of the IP Multimedia Subsystem (IMS) services. Push To View (PTV) A service that allows images to be shared within a closed user group at the touch of a button. PTV is one of the IP Multimedia Subsystem (IMS) services.
Sufficient bandwidth to carry the required traffic. Quality of Service (QoS) and resource admission control to ensure a good user experience (low cell loss ratio, limited jitter, etc). It should be noted that in a mobile environment, the broadcast mode allows the same content to be distributed to
95
www.alcatel.com/atr
all users in the same coverage area, thereby compensating for the limited frequency availability. Gaming services This type of application involves many people interacting with one another via a central server. Access network requirements include: Good interactivity-oriented QoS (round trip delay, cell loss ratio, traffic priority). Rapid handover in the same way as real-time voice/video communications. Seamless handover of an ongoing multi-player gaming session is of paramount importance. A player involved in a game doesnt want to suffer any disruption while playing, or have to re-enter the game (at the risk of disturbing other players) when moving from place to place. Corporate VPN services: access to enterprise intranet/email This group of applications offers enterprise users exactly the same services as detailed above for residential users. The constraints to be addressed by the access networks include:
Access (HSDPA), or CDMA2000. Currently, the main applications used in such situations are voice calls, messaging and corporate VPN services. However, broadcast and gaming applications could soon become more important. The third situation for the user is full mobile usage, using any available mobile access network technologies, such as Enhanced Data rates for GSM Evolution (EDGE), UMTS / HSDPA and CDMA2000. Today voice is still the predominant service, followed by data messaging (SMS / MMS / email). However, there is a clear trend towards the increased use of data, video and gaming services.
Offer from the Operators

Operators can provide user-centric broadband services to subscribers in many ways, the best solution being based on the operators current situation. Two main types of operator can be identified: vertically integrated operators controlling the different accesses (fixed, mobile, etc) and horizontally integrated operators offering services via roaming agreements or in close partnership with other operators. Vertically integrated operators are characterized by the fact that they control and offer converged services via the different accesses they control. Such operators can be further subdivided into: Incumbent operators that own both fixed and mobile networks and are becoming full service providers by integrating their fixed and mobile operations. They offer a mix of their own basic set of services together with a variety of additional rich services from partners. Fixed operators (competitive DSL, cable) can also take on the role of Mobile Virtual Network Operators (MVNO) or introduce wireless technology (WiFi, WiMAX) in their networks, enabling them to offer a similar (or the same) set of services as in the previous case. Mobile-only operators can extend their wireless access (coverage, speed) by deploying new wireless technologies, such
Voice and data Virtual Private Networks (VPN) must follow the employee wherever he or she goes. Thus all the access technologies that an employee might use in different situations have to interact to provide this feature. Security, QoS and performance levels must meet the stringent corporate requirements. Handover is mandatory. It should be rapid for established voice/video Figure 3: Typical user situations calls, but is less stringent for data sessions. Various user situations Three main types of user situation encountered are shown in Figure 3: fixed at home or in the office, nomadic in hot spots and hot zones, and mobile elsewhere. The first typical situation is fixed usage, either at home or in the office, using network technologies such as DSL, cable, satellite and fiber (Fiber To The Home, FTTH; Passive Optical Networks, PON; Ethernet) or fixed wireless access technologies, such as WiMAX, using any type of application, such as voice calls, Internet access, messaging, video and online gaming. The second situation is nomadic usage at hotels, airports or railway stations, making use of WiFi, WiMAX or mobile-oriented technologies, such as UMTS High Speed Downlink Packet
96
Fiber, DSL, cable, fiber at the enterprise Fiber, DSL, cable, fiber at the enterprise
WiFi/DSL, cable, Sat at home
Internet
Corporate
AP
AC
Aggregation Network
WiFi hot spots WiMAX hot zone
BTS/Node B
BSC/RNC
2.5/3G RAN
www.alcatel.com/atr
as WiFi and WiMAX. These operators will be limited in the bandwidth they can offer and possibly also in terms of coverage as they dont offer fixed high-speed access. To overcome this weakness, mobile-only operators are starting to offer DSL broadband services. All vertically integrated operators need roaming agreements with network access providers in regions where they dont own the access. These operators can also offer a single subscription, a single bill, a single directory and a single access portal. As horizontally integrated operators do not own different access technologies, they have to rely on (roaming) partners or run their services transparently over anothers network. Apart from the roaming cases, there will be a need for multiple subscriptions resulting in multiple bills and multiple service offers: Fixed operators can offer converged services via an application, such as the Alcatel Intelligent Mobile Redirect (IMR) solution, that directs calls to either a mobile or a fixed network, depending on the users location. However, it will be difficult to offer the same range of services as a vertically integrated full service provider. Nevertheless, evolution to the IP Multimedia Subsystem (IMS) will mean that simple roaming agreements will allow a fixed operator to offer all its services on any network, since IMS is common to fixed (Telecommunications and Internet converged Services and Protocols for Advanced Networking; TISPAN) and mobile (Third Generation Partnership Project; 3GPP) networks. Mobile operators can offer services by providing mobile services via fixed and mobile access, for example, using Unlicensed Mobile Access (UMA), with fixed broadband access used as a pure bit pipe. MVNOs own the service platform and the subscriber base, and are therefore similar to the mobile operators from a service perspective. ISPs and wireless ISPs taking on the role of application service provider may find it more difficult, unless they take on the role of an IMS provider without an access network.
The second solution, which is based on the Session Initiation Protocol (SIP), is IMR. This solution can be deployed by all operators and used by everyone, including large enterprises. IMR and UMA allow a user to be reached on one device, possibly using a single phone number, irrespective of where he or she is located [1]. Both solutions accomplish this by: Automatically selecting the most suitable access network (usually the one offering the cheapest tariff). This might be either a cellular access network or a WiFi / Bluetooth access network, which generally includes a broadband VoIP network. Rerouting incoming calls towards the appropriate access network.
Table 1: Current solutions overview
Solution Control protocol LAN access technology Cellular access technology IMS support UMA UMA WiFi/Bluetooth GSM/GPRS/EDGE Yes IMR SIP WiFi/Bluetooth All (GSM, CDMA, etc) Yes
GPRS: General Packet Radio Service
Typical Implementations
Three typical implementations will illustrate how operators can offer users an always best connected experience. Voice calls using various access networks As shown previously, the main requirement for a conversational service to different access networks is that the user should be reachable when on the move, with the voice VPN following the employee in the corporate case. In an enterprise, phones will combine WiFi, cellular connectivity and VoIP, while at home mobile phones will evolve towards Bluetooth or WiFi / cellular handsets. Two main solutions can be envisaged for making the seamless use of these dual-mode phones a reality: The first solution, based on UMA technology, allows real-time handover between access networks. However, it is mainly suitable for mobile operators (or possibly MVNOs); it can be used by residential users either at home or in public hot spots, or in combination with an (IP) Centrex solution.
Data services Mobile 3G networks offer significantly higher data rates than 2G networks. However, there is a trade-off between data rate and user mobility; a user who is traveling at high speed cannot expect the same high data rate as a user who is standing still. Recently, IP-based wireless technologies (in particular WiFi) have received a strong boost, providing relatively high data rates at low prices. They can complement mobile networks when necessary. This leads to the introduction of multiple access systems (cellular + IP based) within the same coverage area. Interworking between access technologies has already been demonstrated in 3GPP Release 6 with the work item for 3GPP WLAN interworking [1]. Multiple access systems mean new operator and user requirements; the user may wish to influence the selection of the access system based on such considerations as supported QoS, security level, mobility, price and coverage. On the other hand, the network operator might wish to influence access selection by setting policies. It is expected that customers using multiple access systems will require service continuity as they switch from one to another. This means that their sessions will be subjected to only minimal Access-aware interruption. In addition, the services service provided should be made access-aware The service is (e.g. choose video quality based on the adapted according available bandwidth) and homogeneous to the characteristics to enhance the user experience. of the access Mobility between access networks technology being (WLAN, WiMAX, UMTS, etc) could be used (e.g. video coordinated using the Mobile IP (MIP) quality based on protocol. This protocol reinforces the the available operatorcustomer relationship, bandwidth). enabling a mobile data service for users to be provided not only in the
97
www.alcatel.com/atr
operators own network but also in users enterprise networks and even in the access networks of other friendly network operators [2]. Video on mobile Mobile broadcasting is more than the simple extension of TV broadcasting to mobile terminals. It ranges from the delivery of live video content to customized information services using pushand-store mechanisms. Several technologies are competing for a place in the market for these services. They can be categorized into two main trends: Optimization of digital TV broadcast technologies to improve their performance in a mobile environment. Leveraging the complementary nature of fixed, cellular and broadcast technologies to benefit from their respective strengths. Terrestrial and satellite broadcasting networks are evolving by adapting their capabilities to the mobile environment with the development of the Digital Video Broadcast Handheld (DVB-H) and of the Satellite Digital Multimedia Broadcasting (S-DMB) standards. In the meantime, mobile service providers are starting to deliver streaming video on 2G/3G networks, with the objective of migrating towards the more efficient multicasting capabilities offered by the Multimedia Broadcast / Multicast Service (MBMS). A combination of these technologies is seen as the optimum model for the successful introduction of video on mobile by delivering content efficiently, and offering interactivity and personalization.
Conclusion
Today, people want more than unrelated (i.e. standalone) broadband services. User-centric broadband means that customers share a unified experience using a single identity to obtain different services, without considering the access technology (fixed, mobile or broadcast), and without thinking about how to maintain the connection. User-centric broadband means network interoperability, but also interoperability of applications across a wide variety of devices. It enables people to use any device and always be connected to the most suitable network anytime and anywhere. The model changes from the view of killer applications delivered over a single infrastructure to one that embraces the demand for a customized user experience across multiple devices, networks and applications. Service providers will propose personalized services that recognize the unique needs of individuals. They will differentiate their offerings in a competitive marketplace with customized service bundles that can attract new consumers and corporate users, reduce churn and increase revenue. Alcatel provides communication solutions to telecommunication carriers, Internet service providers and enterprises for delivering voice, data and video applications to their customers and employees. Its leading position in fixed and mobile broadband networks, applications and services, will help partners and customers to build a user-centric broadband world.
References
[1] Ph. Lain et al: Unbounded Mobility: Always Connected, Anywhere, Alcatel Technology White Paper, February 2005. [2] I. Gmez Vinagre et al: Mobile Broadcasting: Extending the Mobile Experience with Efficient Content Delivery, Alcatel Technology White Paper, January 2005.
Philippe Lain is Network Strategy Manager in the Alcatel Network Strategy Group, Paris, France. (Philippe.Laine@alcatel.com)
Johan De Vriendt is Strategy Director NGN in the Alcatel Network Strategy Group, Antwerp, Belgium. (Johan.De_Vriendt@alcatel.be)
Louis Le Gouriellec is Networking Director in the Chief Technology Office of the Alcatel Mobile Communications Group, Vlizy, France. (Louis.Le_Gouriellec@alcatel.fr)
98
www.alcatel.com/atr
ANY ACCESS
J-L.Hurel, J. Brouet, L. Le Gouriellec, M. Peruyero
UNIVERSAL BROADBAND ACCESS: GOING WIRELESS AND MOBILE

Are the emerging broadband wireless and mobile technologies a threat or an opportunity? Will they compete or complement one another?
oing broadband wireless and mobile is not a question of if, but how and when? Be it with Universal Mobile Telecommunications System / High Speed Downlink Packet Access (UMTS/HSDPA), WiMAX, Code Division Multiple Access (CDMA), Time Division Synchronous Code Division Multiple Access (TD-SCDMA), UMTS Time Division Duplex (TDD), WiFi or mobile broadcast technologies, there is a market for it! Dont ask what the users would do with a higher bandwidth and improved Quality of Service (QoS)! Is there a need for mobile triple play? The answer is yes! Will many customers want to watch video clips on a small screen? Without doubt, yes! Users are ready to adopt and pay for services that are personalized, interactive, simple and carried over the best access. The radio access network is a masterpiece in the transport of ad hoc services and will demonstrate its flexibility to achieve the most stringent performanceto-cost ratio objectives. It offers multi-access provisioning, the highest data rates, the lowest latency and best QoS in the nomadic and mobile environFigure 1: Benefits of the ments.
All proposed services and solutions meet users needs: - Broadband must be accessible anywhere, in any situation: at home, at the office, outside and inside, on the pause and on the move. - Access to personal broadband services should be easy, whether fixed, nomadic or mobile. - Users want multimedia services and the largest possible bandwidth at attractive prices. Several criteria are crucial when selecting the right technology: - Optimization of coverage, throughput per user, capacity per site and per cell, mobility and nomadicity conditions. - High spectral efficiency solutions that optimize radio resources management, making it possible to increase traffic throughput. - Reliable handover, roaming and security. Access must be cost-effective and maximize use of the operators three major assets: subscriber base, base station sites and spectrum (licensed and unlicensed).
key access technologies
Driving Forces behind Broadband Wireless and Mobile

The value-proposition offered by mobile and wireless operators is based on numerous multimedia services delivered over fixed or mobile networks, or the Internet. Alcatels aim is to provide these operators with all the business and technical tools they require to put the Internet in each pocket using the best broadband wireless technologies. A user-centric broadband world will be built using selected technologies; Alcatels technology-agnostic approach answers the key business, technical and strategic challenges. The key technologies are shown in Figure 1. The main driving forces for the success of broadband wireless and mobile are:
Speed/user
Very High 2-10 Mbit/s High 1-2 Mbit/s Medium 100-900 kbit/s
Broadband Wireless
FTTx xDSL
Broadband Fixed: From DSL to FTTx
WiMAX WiFi TDD HCR
Broadband Mobile: From 3G to B3G
HSDPA UMTS TD-SCDMA EDGE
"Fixed"
"On the pause" Hot spots
"On the pause Always on" Hot zones Scattered coverage
"On the move" Global coverage
Mobility
3G: Third Generation B3G: Beyond Third Generation DSL: Digital Subscriber Line
EDGE: Enhanced Data rates for GSM Evolution HCR: High Chip Rate
www.alcatel.com/atr
99
The total cost of ownership at the access level must follow the general cost reduction trend, in line with devices and handsets, to ensure the widespread penetration of broadband services. To achieve this, various approaches are being used. High re-use of existing base station sites for rapid, low risk deployment. Flexible capacity growth and initial and additional investments that are closely aligned with the growth in capacity. Evolution to take advantage of new technologies in a futuresafe way. Alcatels radio access network solutions are based on three key pillars: 1 Moving from multi-standard to multi-access: A flexible, cost-effective base station architecture allows the deployment of not only GSM/EDGE and UMTS/HSDPA, but also WiMAX, TD-SCDMA, UMTS-TDD and satellite mobile broadcasting. 2 Cost optimization program every nine months to ensure the scalability needed to allow incremental investment in the infrastructure, hardware and software flexibility, and full backward compatibility to maximize the use of earlier investment. 3 Rapid introduction of new technologies via software upgrades. EDGE is being introduced via software activation; once UMTS is deployed, it will be possible to introduce HSDPA by upgrading the software. The same is true for the smart antennas solutions being introduced in base stations. Alcatels view is that a future radio access solution will be based on collaboration between various technologies to serve different market needs as efficiently as possible. Figure 2 shows the portfolio of solutions Alcatel is offering broadband mobile and wireless operators to meet their short- and mid-term objectives. Alcatels vision is based on various radio interfaces around a common network architecture, with full flexibility in the various building blocks, as depicted in Figure 3.
Coverage and the number of existing sites that can be reused to minimize deployment costs. Average and peak throughput per sector to evaluate the system capacity on the air interface and dimension the terrestrial interface that feeds the radio sites. Average and peak throughput per user, as this affects the types of service that can be offered to subscribers. The BWA technologies considered here are EDGE, UMTSFDD (Frequency Division Duplex) with HSDPA, UMTS-TDD with HSDPA, WiMAX and CDMA2000 1xE-DO.
Figure 2: Alcatel access technologies roadmap and evolution
TD-SCDMA CDMA 2000 UMTS FDD HSDPA HSDPA evolutions Convergence
Mobile
Nomadic UMTS TDD HCR Wireless DSL MC-CDMA 2004 WiMAX 802.162004 2005 Core Network Convergence 2006 WiMAX 802.16e Core & Radio Access + Service Bundling
Driving the convergence to provide truly multi-access broadband

MC-CDMA: Multi-Carrier CDMA
Figure 3: Towards a full multi-access platform
The Multi-access Platform GSM/EDGE, WCDMA, TD-SCDMA & WiMAX
Multi-standard cabinet
S T U R M X U S T U R M X U S U T M R U X
Baseband Baseband Baseband
Multi-band RF Front-end Multi-standard Baseband Multi-standard Transceiver
Performance of Broadband Wireless Technologies

The technical characteristics of the various Broadband Wireless Access (BWA) solutions must be thoroughly assessed before one is chosen. Various radio performance indicators are needed as inputs to the economic assessment that identifies the optimum operator strategy:
100
Multi-Standard & Multi-Access Base Station

SUMU: Station Unit Mobile UMTS TRX: Transmitter/Receiver
Reconfigurable Radio Front-end & Baseband Processing using Software Defined Radio & IP Transport
WCDMA: Wideband Code Division Multiple Access
www.alcatel.com/atr
Overview of BWA standards Table 1 gives an overview of the parameters that influence radio performance. The frequency band has a major impact on the cell radius The higher the frequency band, the lower the range (cell radius), which is why high data rate technologies have a smaller maximum reach than that offered by GSM900 systems. The duplex mode defines the way in which bandwidth is shared between the downlink (base station to terminal) and the uplink (terminal to base station). It affects the system capacity and spectrum requirements. In FDD mode, the downlink and uplink use different frequency channels and are adapted to symmetric traffic. FDD requires paired spectrum allocation. In contrast, in TDD mode the uplink and downlink share the same frequency channel in time. This mode is suitable for asym-
Table 1: Main radio parameters for selected BWA systems

GSM-EDGE 2G (850/900/ 1800/1900 MHz) FDD 200 KHz UMTS-FDD (HSDPA) 3G (2.1 GHz) FDD 5 MHz DS+SS+AMC (QPSK/ 16QAM) CDMA UMTS-TDD (HSDPA) 3G+BWA TDD 5 MHz 10 MHz (BWA) DS+SS+AMC (QPSK/ 16QAM) TD-CDMA WiMAX CDMA 2000 (EV DO) 2G+3G+450 MHz FDD 1.25 MHz DS+SS+AMC (QPSK/8PSK/ 16QAM) CDMA
Frequency band Duplex mode Channel bandwidth
BWA (2.5/3.5GHz) TDD/FDD 1.25 to 20 MHz QFDM+AMC (QPSK/16QAM/ 64QAM) TD-OFDMA
Physical layer
AMC (GMSK/8PSK) TDMA 9 (traffic channel), 14 beacon channel 2x4.6 MHz
Access layer Frequency reuse Minimum spectrum
3 30 MHz (for 10 MHz channel)
2x5 MHz
1x5 MHz
2x1.25 MHz
AM: Amplitude Modulation QAM: Quadrature Amplitude Modulation GMSK: Gaussian Minimum Shift Keying
QPSK: Quaternary Phase Shift Keying PSK: Phase Shift Keying
Figure 4: Different multiple access modes
Power
Power and codes
200 KHz 600 KHz Time
Frequency
TDMA (GSM-EDGE)
5 MHz (UMTS FDD) 1.25 MHz (CDMA 2000) Time
Frequency
TD-CDMA (UMTS-TDD)
Power Power and codes

Up lin k
10 to 15 MHz 5 MHz (UMTS FDD) 1.25 MHz (CDMA2000) Time Frequency Time 1.25 to 20 MHz
Frequency
CDMA (UMTS-FDD/CDMA2000)
TD-OFDMA (WiMAX)
Do
nl in
Do w
nl in
www.alcatel.com/atr
Up
lin
101
EDGE throughput can be optimized if EDGE carriers are metric traffic, since usually the ratio between uplink and downdeployed with higher frequency reuse (thanks to the reduction link is adjustable. TDD can be deployed in unpaired and in the level of interference). WiMAX has, by its nature, simipaired spectrum allocations. lar requirements to GSM/EDGE systems in terms of interferChannel bandwidth directly affects the throughput on the ence levels. However, adaptive antenna technology means that air interface. The greater the channel bandwidth, the higher WiMAX can be deployed with a frequency reuse of just three. the data rate, which is why WiMAX systems have much Table 1 shows the spectrum required for deploying a BWA soluhigher throughputs than others. tion, based on frequency reuse, channel bandwidth and The physical layer of any BWA systems is based on Adapduplex mode. tive Modulation and Coding (AMC) mechanisms. This enables the fluctuating propagation channel characteristics to be efficiently exploited by selecting higher level modulation schemes Radio throughputs of BWA technologies when possible to increase the throughput per sector. In addiThe results synthesized in Figure 5 assume that all the radio tion, the modulation technology affects the performance. network resources are being utilized; the comparisons were CDMA (UMTS and CDMA2000) use Direct Sequence Spread made using the same assumptions about the radio environments Spectrum (DSSS): narrowband signals are spread over a (e.g. same propagation conditions, same antenna height, larger bandwidth signal, which is more robust against interfersame indoor penetration requirements). The average throughence and has improved sensitivity (processing gain). puts can then be seen as the minimum achievable throughputs. GSM/EDGE and CDMA systems use single carrier modulation. In addition, a downlink / uplink ratio of 3:1 is assumed for TDD In contrast, WiMAX is based on Orthogonal Frequency Division systems (UMTS-TDD and WiMAX). Multiplexing (OFDM), which is a multiple carrier modulation In terms of throughput per sector, three performance system. The high data rate information flow is transmitted in groups can be derived: parallel on a higher number of orthogonal narrowband subcarriers (512, 1024 or 2048). OFDM offers the best performance 300 kbit/s: 2G technology (GSM/EDGE) can offer 320 kbit/s / complexity tradeoff for transmission bandwidths larger than per carrier on average 5 MHz, making it one of the main building blocks for fourth gen 700 kbit/s to 2 Mbit/s: 3G technologies enhanced with eration (4G) systems. UMTS/HSDPA or CDMA2000 1xEV-DO provide average data The multiple access scheme (see Figure 4) indicates how rates per sector of about 1 Mbit/s (from less than 1 Mbit/s for the available bandwidth is shared between users (which CDMA2000 1xEV-DO to around 2 Mbit/s for UMTS/HSDPA). impacts the throughput per user) and how the system could In these systems, the ratio between the average and peak be deployed (frequency reuse and average throughput per sector). In the case of GSM/EDGE using Time DiviFigure 5: Typical downlink throughputs sion Multiple Access (TDMA), the user data is divided between timeslots belonging to a given channel; the user WiMAX (10 MHz) data can be allocated to a maximum of WiMAX (5 MHz) four timeslots per TDMA frame; each timeslot can carry different user data. CDMA2000 EV-DO Average However, in UMTS-FDD and Peak CDMA2000 systems, which are based UMTS TDD (5 MHz) on CDMA, users share the whole sysThroughput tem bandwidth and are allocated difHSDPA FDD (5 MHz) per sector ferent codes and powers. The use of EDGE 900 (200 MHz) codes enables CDMA systems to operate with a high level of interference. 10 100 1000 10000 100000 Orthogonal Frequency Division Multiple Access (OFDMA) is used in Data rate [kbit/s] WiMAX systems. Frequency reuse defines the minimum number of frequency blocks WiMAX (10 MHz) that are required for cellular deployments in BWA systems. Indeed, since BWA systems use AMC, the throughput per sector depends on the level of interference created by the cells transmitting at the same frequency. In CDMA systems, a frequency reuse of 1 is common (only one paired frequency block is required to deploy a CDMA system). Of course, this generates a high level of interference across the cell. In the case of GSM/EDGE, a reuse of at least nine is required; the
102
WiMAX (5 MHz) CDMA2000 EV-DO UMTS TDD (5 MHz) HSDPA FDD (5 MHz) EDGE 900 (200 MHz) 10 100 1000 10000 100000
Average Peak Throughput per user
Data rate [kbit/s]
www.alcatel.com/atr
rates is rather low because of the high level of interference across the cell resulting from a frequency reuse of 1. > 10 Mbit/s: WiMAX offers average data rates of around 15 Mbit/s, with peaks up to 22 Mbit/s. Considering the available data rates per subscriber, the various BWA options support Asymmetric Digital Subscriber Line (ADSL) like services at the following speeds: Up to 128 kbit/s for GSM/EDGE. Up to 512 bit/s for 3G technologies (data rate at the cell edge is lower). Up to several Mbit/s for WiMAX. Coverage and site reuse Coverage determines the number of sites that are required to serve the entire service area. It is thus of the utmost importance when considering the investment needed to introduce BWA. Figure 6 shows the cell ranges for dense urban and rural environments with the following constraints: Deep indoor penetration, which is essential for BWA systems. Minimum uplink data rate per user at cell edge of 64 kbit/s.
Figure 6: Typical range (uplink data rate 64 kbit/s at the cell edge)
WiMAX (RG) WiMAX (PCMCIA) CDMA2000 (2 GHz) UMTS TDD (3.5 GHz) UMTS TDD (2 GHz) HSDPA FDD (2 GHz) EDGE 1800 EDGE 900 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Typical Range Dense Urban
Range [km]
WiMAX (RG) WiMAX (PCMCIA) CDMA2000 (2 GHz) UMTS TDD (3.5 GHz) UMTS TDD (2 GHz) HSDPA FDD (2 GHz) EDGE 1800 EDGE 900 0 1 2 3 4 5 6 7 8 9
Typical Range Rural
Range [km]
Figure 7: Performance metrics for BWA technologies
>
D. Renaudeau, D. Boettle, H. Steyaert: WiMAX: From Fixed Wireless Access to Internet in the Pocket, Alcatel Telecommunications Review, 2nd Quarter 2005, pp 144-149 (this issue).
2G Site Reuse Urban

1
0,1
Two scenarios are analyzed for the Frequency WiMAX system. WiMAX PCMCIA correReuse sponds to Internet in the Pocket, the terminal being a PCMCIA card that can be plugged into a laptop or a Personal Digital Assistant (PDA). The WiMAX residential gateway corresponds to a deployment targeting DSL-like wireless with the terminal having an ADSL modem form factor, including a small embedded antenna, and hence capable Maximum of higher transmission powers. Throughput Figure 6 clearly shows that the per Sector higher the frequency, the lower the range. A 900 MHz GSM/EDGE system provides the best coverage; EDGE coverage is the same as GSM coverage. For frequency bands of around 2 GHz and above, the range is at least halved compared with EDGE at 900 MHz. However, the ranges cited here are maximum ranges for GSM; in practice, GSM/EDGE ranges are limited by their capacity (typ-
2G Site Reuse Rural

0,01
EDGE HSDFA-FDD CDMA2000 WiMAX RG WiMAX PCMCIA Average Throughput per Sector ically 400 m in dense urban areas, 7 km in rural areas). Consequently, in dense urban areas, 2G sites can be used to deploy any BWA technology while providing complete coverage for BWA services. This is a key feature considering the difficulty of finding additional sites in urban areas.
103
www.alcatel.com/atr
Comparative view Figure 7 compares the performances of GSM/EDGE, UMTS-HSDPA, CDMA2000 and WiMAX; the following conclusions can be reached:
Figure 8: Rural area cases: What technology best meets the service requirements
Data peak rate

512 kbit/s Data Centric 100% data 15% voice
W CDMA W
WiMAX@3500
W W
GSM/EDGE enables BWA to be introduced smoothly at low cost: no addi256 kbit/s tional sites are required, thereby optimizing capital expenditure and minMarketing Offers imizing operating expenses. EDGE can be introduced using a simple 128 kbit/s software upgrade; Alcatels EvoliumTM Voice Centric hardware has been EDGE ready since 15% data 2001. Services requiring data rates of 100% voice up to 128 kbit/s can be introduced. 45 kbit/s UMTS/HSDPA technology to support more users and higher data rates for BWA services in urban areas. Alcatel EvoliumTM multi-standard base stations make optimum use of earlier 2G investments by reusing the same cabinet for both GSM and UMTS. Users receive data throughputs of up to 512 kbit/s. WiMAX is a real Internet in the Pocket BWA solution, offering high data rates per sector, and data rates in excess of 1 Mbit/s per user. Alternatively, WiMAX can be viewed as a very high capacity solution for low data rate user services, or as an overlay fixed wireless access (using WiMAX PCMCIA). The performance results outlined here are fed into the economic models used to select the most suitable technology based on the target market, the target area and the shortest return on investment period.
CDMA
W CDMA
CDMA
CDMA EDGE
EDGE CDMA
EDGE CDMA
CDMA@450
CDMA CDMA EDGE EDGE CDMA EDGE CDMA
EDGE@900 2/km2 7/km2 13/km2 20/km2
User Density
The following results show that both groups of wireless technologies offer specific benefits that are needed to offer a universal wireless service in any situation (full mobile, nomadic, fixed), in any area (rural or urban), at different bitrates, and whatever the subscriber density. The overall finding is that none of these technologies can provide the optimum solution for all the scenarios, hence they should to be used to complement one another, taking the best features from each (see Figure 8). WiMAX, EDGE and CDMA450A as fixed wireless access technologies for rural areas These three wireless technologies were compared to assess how they perform economically in different density zones in a rural area while supporting the required voice + data capability. The results show that all three technologies can be used, taking full advantage of the benefits of each. However, each solution has its limitations: WiMAX is not yet suited to voice-oriented services, mainly for handset availability reasons. Moreover (at least in the short term), because of its mobility limitations it cannot support a full voice service. GSM/EDGE and CDMA hardly achieve the data capacity needed to support high-speed Internet access. Nevertheless, each solution has its advantages: CDMA 450 is the most economic answer in very low density rural areas thanks to its unrivalled geographical reach. It fully supports the voice service, while offering an acceptable data rate (comparable to the lowest class of fixed DSL). It could be deployed with a WiMAX implementation so that high bitrates can be offered where the user density exceeds between two and seven subscribers per square kilometer. EDGE and CDMA 450 are profitable in rural areas with medium to high population densities, with data rates limited to 128 kbit/s. It could be complemented by a WiMAX solution to provide higher data rates.
Key Broadband Wireless Technologies for Rural and Urban Environments

To help understand the positioning of the different solutions, two wireless technology groups are compared for both rural and urban areas: first a data-oriented group using WiMAX, and second a mobile-oriented group featuring EDGE, CDMA2000 1xEV-DO using 450 MHz in rural areas (CDMA450) and UMTS/HSDPA CDMA2000 1xEV-DO in urban areas. Two dimensions have been used to highlight the differences in profitability between the technologies: Variations in density: Four cases for rural fixed wireless access and three urban cases: dense urban, urban and suburban. Variations in data rate requirements corresponding to different marketing packages; all include both voice and data: - Rural case: Four data rates corresponding to two voice-centric marketing offers (45 and 128 kbit/s) and two data-centric offers (356 and 512 kbit/s). - Urban case: Three data rates corresponding to three well known segments: a full mobile offer up to 512 kbit/s; a nomadic / hot-zone offer up to 1 Mbit/s, and a fixed wireless offer up to 2 Mbit/s. Note that regular design rules and classical business case rules were used to quantify the equipment needed and the profitability for each (density + data rate) case.
104
www.alcatel.com/atr
WiMAX is the best answer for data rates of up to 512 Mbit/s. This is no surprise as it is a data-built-in wireless technology. It can be deployed in overlay mode on top of either EDGE or CDMA, thus offering future-safe evolution. Increased demand for voice and data services could be met either using more WiMAX + EDGE or CDMA density extension, or simply by increasing WiMAX density, using Voice over Internet Protocol (VoIP) to carry voice as soon as terminals become available. WiMAX, UMTS/HSDPA and CDMA2000 1XEV-DO wireless technologies in urban areas The two groups of wireless technologies were compared to assess how they perform economically in the different density zones of an urban area while supporting the required voice + broadband data capacity. The results (see Figure 9) show that all three technologies can be used, taking advantage of their particular benefits.
Figure 9: Urban area cases: comparing service requirements
usage flexibility, thereby enabling the existing mobile infrastructure and frequency resources to be reused when adding mobile / moderate nomadic broadband data services. WiMAX appears to be the best broadband data technology for any operator thanks to its DSL-like capacity, in all data-oriented cases, as already observed in the rural case. Consequently, it is the natural complement to HSDPA or EV-DO for operators that want to address all the mobile, nomadic and fixed wireless segments. WiMAX can be used for fixed wireless DSL application in areas with limited copper outside plant, as well as to provide a nomadic / hot zone wireless broadband service with HSDPA or EV-DO to meet the need for full mobility.
Conclusion
Several broadband wireless access technologies are available for different uses, providing different performances and suited to different geographies. In all cases, the aim is to offer the best access network for users when and where needed. Alcatel can help broadband and mobile service providers to choose the best mix of these technologies as WiMAX there is no one fits all solution; it is essential that the technologies should complement one another. Alcatel offers three main advantages:
WiMAX HSDPA EV-DO
Data rate
WiMAX
DSL-like DATA rate Home or office Marketing Case
2 Mbit/s
WiMAX
WiMAX
High DATA rate Nomadic, indoor/outdoor
1 Mbit/s
WiMAX HSDPA EV-DO
WiMAX HSDPA EV-DO
Moderate DATA rate Full mobile
HSDPA/EV-DO 512 kbit/s

HSDPA EV-DO HSDPA EV-DO HSDPA EV-DO
Suburban:100
Urban:250
Dense Urban:500
Density Subscribers/km2 Economically speaking, the results show that all the technologies have limitations: WiMAX is not yet ready for a full mobile offer, primarily because of a lack of suitable handsets. HSDPA and CDMA EV-DO both achieve only limited profitability when addressing nomadic hot zones. Using them to offer a combined mobile data / nomadic package could be problematic, as the larger number of sites needed would adversely affect the business case. Other drawbacks are that more equipment (and therefore investment) is needed as the data rate increases (e.g. 2 Mbit/s instead of 1 Mbit/s for nomadic use would dramatically affect the business case), and that the peak rate at the cell edges is only 128 kbit/s, compared with around 4 Mbit/s for WiMAX. Again, each solution has specific benefits, so they can be used to complement one another: UMTS-HSDPA and CDMA 1xEV-DO are definitely the best fit for full mobile services, and can support a healthy throughput of 512 kbit/s on top of regular voice. They also offer greater
Its leading position in fixed and mobile broadband networks, applications and services, particularly its worldwide leadership in DSL. Its future-proof radio access network solutions which enable major new technologies, such as HSDPA, to be deployed by upgrading the UMTS software. Its strong commitment to optimum nomadic usage and service ubiquity combining open technology assessment, strong partnerships and network integration for the various wireless access technologies, including WiFi, WiMAX, UMTS TDD and mobile broadcast technologies.
Jean-Louis Hurel was Product Marketing and Business Office Director in the Alcatel Mobile Radio Division until April 2005 and is now 2G Product Line Director in the same Division, Vlizy, France. (jean-louis.hurel@alcatel.fr) Jrme Brouet is Network Design Manager in the Alcatel Mobile Radio Division, Mobile Communications Group, Vlizy, France. (Jrome.brouet@alcatel.fr)
Louis Le Gouriellec is Networking Director in the Chief Technology Office of the Alcatel Mobile Communications Group, Vlizy, France. (Louis.Le_Gouriellec@alcatel.fr) Michel Peruyero is Executive Director in the Chief Technology Office of the Alcatel Mobile Communications Group, Vlizy, France. (Michel.peruyero@alcatel.fr)
www.alcatel.com/atr
105
TECHNICAL PAPER
ANY ACCESS
B. Haberland, W. Koenig, A. Pascht, U. Weiss
SOFTWARE DEFINED RADIO: A PROMISING TECHNOLOGY FOR MULTI-STANDARD BASE STATIONS

SDR is an important step towards offering every user at any location the best service, bandwidth and mobility, in line with Alcatels vision of a user-centric broadband world.
ireless systems are becoming more and more heterogeneous in terms of standards and the frequency bands they use. In the case of the Base Station Subsystem (BSS) and Radio Network Subsystem (RNS), the reusability of existing sites is particularly important when it comes to introducing a new radio standard because it has significant benefits for operators in terms of savings in both capital expenditure and operating expenses. Multi-standard multi-band solutions are essential to enable equipment suppliers to reduce the number of development cycles, thereby cutting R&D costs and shortening the Hot spot time-to-market. WLAN/WiMAX access point made Fewer variants publicly available mean much by a service lower mainteprovider. Refers to nance, manufaca single cell turing and modenvironment. ule improvement (cost reduction) Hot zone efforts. However, An assembly of all these advanseveral hot spots tages have to covering a larger compensate for area. Potentially the possibly supports multi-hop slightly higher connections. cost of Software Defined Radio (SDR) solutions compared with dedicated solutions. In beyond 3G systems, new air interfaces (e.g. WiMAX for hot spots and hot zones) and an extension of the multi-standard radio resource management procedures, are making SDR solutions ever more relevant for base station platforms. For the operator, SDR offers greater flexibility in the deployment of different
106
standards and multiple frequency bands. For example, it can be used to rapidly introduce hot spots / hot zones reusing existing sites.
Basic Principles of Multi-Band/ Multi-Standard Transceivers

A multi-band transceiver (TRX) is a radio front-end that can be operated in a number of frequency bands. This is interesting in conjunction with UMTS Frequency Division Duplexing (FDD), where
Implemented multi-band front-end demonstrator in operation with signal source, analyzer and control environment
Technical Motivation
Alcatels SDR technologies are focusing on multi-band, multi-standard solutions. There is a differentiation between: Multi-band, single standard: - 1.8/2.1/2.6 GHz: Universal Mobile Telecommunications System (UMTS) R99. - 2.5/3.5 GHz: WiMAX according to the IEEE 802.16e standard. Multi-standard within one frequency band: - 2.1 GHz: UMTS R99, High Speed Downlink Packet Access (HSDPA), HSDPA enhancements, High Speed Uplink Packet Access (HSUPA) and the introduction of Orthogonal Frequency Division Multiplex (OFDM) in Third Generation (3G) systems. Multi-band, multi-standard: - 1.8/2.1/2.6 GHz: UMTS R99, HSDPA evolution, 3G OFDM. - 1.8/2.1/2.6 GHz / 3.5 GHz: UMTS R99, HSDPA evolution, 3G OFDM, WiMAX.
the standard already defines a set of operating bands. An Alcatel research project has successfully developed a first demonstration model of a Multi-Band Front-End (MBFE) operating in selected In addition, there is a work item in the frequency bands (see Table 1). Third Generation Partnership Project The frequency bands can easily be (3GPP) for UMTS900 that will pose addi- changed without modifying the hardtional challenges for SDR technology. ware, giving network operators considerable flexibility not only Table 1: Selected frequency bands for MBFE implementation when introducing new frequency Operating Band Receive Frequency Transmit Frequency bands, but also I 1920 1980 MHz 2110 2170 MHz for subsequent III 1710 1785 MHz 1805 1880 MHz network reconIV 1710 1755 MHz 2110 2155 MHz figuration and Extension Band 2500 2570 MHz 2620 2690 MHz optimization.
www.alcatel.com/atr
Figure 1: Overall architecture of the MBFE
MBFE
LNA
Receiver Front-End
To BB processing
To/from O&M
ACE
Control
To/from local control
Power Amplifier
Transmitter Front-End
From BB processing
BB: BaseBand
O&M: Operations and Maintenance
Figure 2: Implementation of (a) transmitter front-end, and (b) receiver front-end
a)
Transmit Front-End
LPF Gain Imbalance Compensation Gain Imbalance Compensation Offset Compensation +HBF +HBF A Offset Compensation Delay Compensation + Input Selection D
Clipping
BTI (I/Q) Debug
0+Phi 90+Phi NCO 0 90 +
LO 0 LPF 90 +
VGA PA
+HBF +HBF DAC
DFE AFE: Analog Front-End BTI(I/Q): Baseband to Transmitter Interface, delivering I and Q data DAC: DigitalAnalog Converter DFE: Digital Front-End HBF: Half Band Filter
AFE
LPF: Low Pass Filter PA: Power Amplifier Phi: Angle (for I/Q Phase Imbalance Compensation) VGA: Variable Gain Amplifier
mode or low IF mode, depending on the frequency of the Numerical Control Oscillator (NCO) in the digital front-end. When the NCO is set to 0 (zero IF mode), the carrier is located around frequency zero in the baseband (Figure 3a, left side) and therefore only a single carrier can be treated in this case. After conversion (mixing) to Radio Frequency (RF) (Figure 3a, right side), in the RF band the carrier then appears symmetrically to the Local Oscillator (LO) frequency set in the analog front-end section (see Figure 3a). In preparation for multi-carrier usage, the carrier can also be separated from the center frequency at a low IF, determined by the NCO frequency (see Figure 3b, left side). In this case, the carrier in the RF band (see Figure 3b, right side) appears to the left or right of the LO frequency. Emissions caused by imperfections in the analog In-phase and Quadrature (IQ) modulator (image and LO leakage) have to be eliminated (i.e. compensated). The extension of the MBFE to a multistandard base station covering UMTS FDD / HSDPA and WiMAX introduces new requirements. The multi-standard, multiband transmitter/receiver needs to handle additional frequency bands and an enlarged bandwidth of 20 MHz, as well as to meet more stringent performance requirements resulting from the need for WiMAX to support OFDM signals. Enabling components Many of the components used to build an MBFE are operated at the edge of what is feasible today. In some areas this will entail significant improvements and developments by the component manufacturers: Mixers: Highly linear IQ (de)modulators working over a wide frequency range are needed to implement the direct-conversion architecture. The first products are just appearing on the market. As part of a nationally funded research project, Alcatel and its partners Lucent and Infineon, have specified and manufactured early samples that meet the requirements. Synthesizer: Wideband synthesizers covering the required frequency range are not yet available. Special samples have been developed for the demonstrator. Converters: Commercial digital-analog and analog-digital converters capable of handling arbitrary carriers inside one band are available. Power amplifier: Wideband applications are impossible using conventional
107
b)
Receive Front-End
Buffer LPF/AAF Data Format Adaptation Delay Compensation DC-Offset Correction Output Selection D CIC+ RCC+ Digital Gain Control A VGA LO 0 90 Buffer LPF/AAF A D IQ Imbalance Correction Carrier Selection
Baseband (I/Q) Data Logger
LAN
CIC+ RCC+
AFE AAF: Anti-Aliasing Filter ADC: AnalogDigital Conversion
ADC
DFE CIC: Cascaded Integrator-Comb filter RRC: Root Raised Cosine filter
Architecture of a multi-band transceiver Figure 1 depicts the modules deployed [1] in the MBFE, each of which must meet much more stringent requirements than those for a single-band approach. The power amplifier (for the selected bands) has to operate over a wide frequency range from 1800 to 2700 MHz. The Antenna Coupling Equipment (ACE) selects the required bands (to and from the antenna) and separates the corre-
sponding transmit (TX) and receive (RX) sections (duplexer). The Low Noise Amplifier (LNA) performs the initial amplification of the received signals. The entire operation of the MBFE is managed by the control module. The transmit and receive front-ends are both implemented using a direct conversion architecture (see Figure 2). The transmit front-end can be (re-)configured for two different modes of operation: zero Intermediate Frequency (IF)
www.alcatel.com/atr
high power devices. However, new technologies (e.g. GaN) are emerging, but so far the achievable power is limited. Multi-band duplexers: Only dual-band duplexers are currently available. Today, the multi-band duplexer is a limiting factor because of its size and price. However, new technologies, like ceramic multi-mode filters, will help to reduce the size considerably. Pre-compensation of dirty RF Wherever analog components are used, it is essential to take account of their imperfections that affect the overall behavior. Many of these imperfections can be compensated using digital pre-processing. Image suppression and LO leakage suppression in the transmitter When operating in low IF mode to meet the spectrum emission mask requirements, the image and LO leakage have to be carefully suppressed. The image is a result of gain and phase imbalances in the I and Q (in phase and quadrature) branches of the analog IQ modulator, and can therefore be compensated by setting the corresponding gain and phase values appropriately in the digital part of these branches. LO leakage results from unpredicted DC offsets in the I and Q paths. In this case, compensation can be achieved by adding the inverse of the DC offset in the digital domain of each path. The values for the compensation parameters are determined by analyzing the analog output signal with the help of a feedback loop.
Figure 3: Spectrum of direct upconversion (a) from zero IF, and (b) from low IF
a) possible frequency channel positions wanted channel (rest of) LO image
f=0
b) possible frequency channel positions wanted channel (rest of) LO image
fRF = fLO
wanted channel
f=0
fRFC = fLO be driven to the nonlinear part of its transfer characteristic, leading to an additional improvement in efficiency. To adjust the pre-distortion parameters adaptively, the output signal of the power amplifier has to be compared with the input signal of the pre-distortion [2]. Figure 5 shows the spectra of a one- and a four-carrier signal with and without digital pre-distortion. Lab test results In the receive chain, with regard to IQ demodulation and analog/digital conversion, it has been proven that the reference sensitivity level can be achieved over all the bands [1]. The current MBFE lab demonstrator comprises up-conversion, including all the reconfiguration and compensation mechanisms plus the wideband power amplifier driver stages. In zero IF mode operation, the specifications can be met across all bands. In low IF mode operation, the compensation mechanisms for image and LO suppression are working well to meet all the requirements, including the spectrum emission mask, which is the most challenging in this case (see Figure 6).
Normalized Signal Magnitude
Peak-to-average reduction This enables the power amplifier to be operated in a more efficient region. Because of the Quaternary Phase Shift Keying (QPSK) or Quadrature Amplitude Modulation (QAM) scheme used in Wideband Code Division Multiple Access (WCDMA) and OFDM systems, the generated multi-carrier signal can feature a high peak-to-average power ratio of more than 10 dB. Without additional measures, the power amplifier would have to be operated with a high back-off from its 1 dB compression point in order to avoid large inter-modulation products. The peak-to-average reduction shown in Figure 4 enables the peaks to be reduced, leading to a peak-toaverage ratio of around 5 dB. The peak-toFigure 4: Peak power reduction average reduction must not generate 3GPP Test Model 3 with 32 Dedicated Physical Channels spectral components outside the Original Signal channel and the Signal with Reduced Peak Power 0.6 error vector magnitude of the result0.5 ing signal must stay within the 0.4 3GPP specification.
0.3 0.2 0.1 0 8.4
8.45
8.5
8.55
8.6
8.65
8.7
8.75
8.8
Samples in Time Domain
Power amplifier linearization The signal data is pre-distorted with a nonlinear transfer characteristic inverse to the nonlinear transfer characteristic of the power amplifier. Consequently, the power amplifier can
Architecture of a MultiStandard Baseband Platform

A multi-standard base station consists of a universal hardware platform that can be configured for a specific air interface by downloading the appropriate functional software. This should be possible during operation without any downtime of the base station. The functionality of the downloaded software components ranges from protocol stacks to physical layer processing algorithms.
108
www.alcatel.com/atr
A multi-standard baseband platform must fit with the requirements of the applications to be supported (radio standards, operating modes) and has to dynamically support varying traffic mixes. Compared with conventional solutions, clear advantages can be obtained in terms of more efficient use of the existing hardware resources and the attainable Quality of Service (QoS). High granularity is needed for the resources allocated to downlink and uplink signal processing. This requires architectural enhancements and intelligent and dynamic hardware resource management. A number of Digital Signal Processors (DSP) and hardware accelerator blocks are used for high bitrate and high volume data processing, as required for bit stream oriented functions. These elements are interconnected via a high speed data and control bus, and supervised by a control system (general purpose processor) (see Figure 7). SDR baseband software library A software library provides the signal processing building blocks needed to generate standard specific processing chains
Figure 5: Spectra of (top) a one-carrier and (bottom) a four-carrier signal (red) without and (blue) with pre-distortion.
Spectral Power Density (dBm/Hz)
20 15 10 5 0 -5 -10 -15 -20 -25 -30 -35 -40 -45 -50 -55
2.120 2.125 2.130 2.135 2.140 2.145 2.150 2.155 2.160
Frequency (GHz)
Spectral Power Density (dBm/Hz)
20 10 0 -10 -20 -30 -40 -50
High granularity Refers to the degree of flexibility regarding the allocation of software tasks to processing elements like DSPs and FPGAs. In the context of re-configuration and load-sharing between different air interface standards it defines how finely the available processing performance can be split and assigned to different air interface functions.
Frequency (GHz)
Figure 6: Direct conversion operated in low IF mode (a) without compensation and (b) with compensation
a)
b)
www.alcatel.com/atr
109
for downloading the downlink and uplink radio functionality to the processing elements, as required by the actual configuration. A suitable real-time operating system on the DSP level provides low layer functions to operate the hardware elements, for example, task management and scheduling. Resource and load management functions are realized using a hierarchical approach from the control system. Before the runtime software is downloaded to the processing elements it is parameterized and linked together according to the required functionality. A scheduler in the DSP then calls the processing blocks in the appropriate way. The processing chains (UMTS or WiMAX entity, etc) can be composed and configured from the dedicated part, the common part and the common algo-
Figure 7: Baseband hardware architecture rithm part of the system library, depending on the Controller Downlink/Uplink required air interData/Control bus face protocol ... (Figure 8). Dedicated system libraries provide the unique ... DSP DSP DSP FPGA FPGA functions for a specific air interface environment, FPGA: Field Programmable Gate Array defined by the given standard (e.g. transport block concatenation, match the funcinterleaving, radio frame segmentation). tionality of a parFast Fourier Common system libraries consist of sys- ticular standard. Transform (FFT) tem functions that are common to different Functions to be Computationally efficient algorithm air interfaces, which may need to be config- considered here for performing a ured by setting suitable parameters to include, for examcomplete discrete ple, cyclic redunFourier transform, dancy checking, i.e. a transform of a spreading, moduFigure 8: Hierarchical SDR library concept for baseband signal processing signal from the time lation and power domain into the control. UMTS WiMax MAC Scrambling Channel EST. scheduling IOTA frequency domain. Common algorithm libraries generally consist of basic functions that are widely used in the WiMax Dedicated OFDM HSDPA UMTS WiMax Entity scientific and telecom fields, such as filterSystem part 3G Library Library Library ing functions, complex/real Fast Fourier Transform (FFT) functions and vector Spreading, maths functions. Turbo Code, The application of a hierarchical SDR library concept has the potential to proCommon vide cost-effective solutions not only for System part software development, but also with inc. Parameterization Common Filtering respect to the memory required to store UMTS Library Functions Entity the huge volume of software codes. However, further studies and evaluations are Common necessary to prove the practicality of this Algorithm part approach [3]. inc. Parameterization DSP
Library
Figure 9: SDR will merge todays separate implementations into a single future-proof module
Impact of SDR Technologies on the Node B Product Portfolio

Alcatel base station Today dedicated solutions are available for different standards (see Figure 9). SDR technologies will enable the processing units to become more flexible in their treatment of different standards and frequency bands, thereby minimizing the number of hardware variants. In addition, it allows dynamic reconfiguration and parallel operation of different standards on the same platform at the same time. Thus it improves utilization of the processing power and therefore the overall system performance. It is clear that multi-standard O&M and transport termination solutions are needed for the Station Unit Module Universal (SUMU). These are realized by software. The introduction of IP transport makes it easier to share the Iub transport
BBMA BBMC SUMU NEMO
TEU ANRU Twin/TEUC ANRU WiMAX
SDR MSBB
SDR MSBB
Transport
BasebandPlatform
Transceiver Front-ends
ANRU: Antenna Network and Receiver UMTS BBMA: BaseBand Module revision A BBMC: BaseBand Module revision C
NEMO: Network and Modem MSBB: Multi-Standard BaseBand TEU: Transmitter Equipment UMTS
110
www.alcatel.com/atr
resources than was the case with the Asynchronous Transfer Mode (ATM). Baseband processing is done primarily with programmable software devices, such as DSPs and hardware accelerators (e.g. field programmable gate arrays). This is an opportunity to introduce multistandard solutions. Shifting most of the complexity of the transceiver architecture (e.g. TEU; ANRU) to the digital part, enables operation in different bands and to different standards. This also has a positive effect on costs, because the price of digital processing components is approximately halving every year, whereas the prices for analog RF components are declining at no more than 10% per year. Evolutionary phases Different evolutionary phases with increasing complexity and flexibility for the operator can be identified for both the baseband part and the transceiver part. The individual steps are described below. Evolutionary phases for the baseband part An example based on a system with a pool of 11 baseband boards demonstrates the different levels of reconfiguration: Reconfigurable software solutions that enable the hardware platform (i.e. all equipped baseband boards together) to be used for either standard A or standard B.
BB platform Standard A BB platform Standard B
Parallel operation of standards A and B with dynamic load sharing at the baseband board level on a call-by-call basis.
Load sharing example 1 Load sharing example 2
Multi-band/multi-standard transceiver Single standard transmitter/receiver with multi-band operation. The following applications are of added value: - 1.7 GHz to 2.7 GHz for 3G+ (UMTS 1800, UMTS core band, UMTS 2600); - 2.5 GHz to 3.5 GHz for WiMAX. Software reconfigurable multi-band/ multi-standard transmitter/receiver (only standard A or B in operation). The following applications can be identified: - 1.7 GHz to 2.7 GHz for 3G+ and WiMAX (2500); - 2.5 GHz to 3.5 GHz for 3G+ (UMTS 2600) and WiMAX (3500). Multi-band/multi-standard transmitter/ receiver in parallel operation (standards A and B operate in parallel). The same combinations as in the previous case are of interest.
figurable base stations. They will enable Alcatel to extend its product portfolio. This is especially true for a multi-standard combination of 3G evolution and WiMAX. SDR technology reduces the number of dedicated developments for different standards and frequency bands. This is important for Alcatel because both fewer product variants and fewer development cycles reduce R&D costs. In addition, SDR-based solutions are future-safe as features can be upgraded by software. For customers, the SDR approach offers new opportunities for dynamic network reconfiguration (e.g. introduction of hot spots) and the reuse of existing sites and equipment, leading to significant reductions in capital expenditure and operating expenses. Furthermore, the high flexibility can serve as the basis for future infrastructure sharing.
Conclusion
Software defined radio is an enabling technology for multi-standard/multiband base stations. A multi-band frontend lab demonstrator has already been built and is operating not only in the UMTS core band but also in the GSM1800 band and the 2.6 GHz extension band. A direct conversion architecture has been used with image suppression and LO leakage suppression. Additional features, including peak-to-average reduction and digital pre-distortion for power amplifier linearization, are also ready to be used. Regarding the multi-standard capabilities of the baseband platform, the hierarchical SDR library concept is a powerful approach to the implementation of cost-effective solutions. Multi-standard/multi-band solutions are relevant for flexible, software recon-
Bernd Haberland is head of all radio activities in the Research & Innovation center, Stuttgart, Germany. He is also project leader of the Alcatel Radio Protocols and BTS Nodes project with activities in Stuttgart and Shanghai, and a Fellow Member of the Alcatel Technical Academy. (Bernd.Haberland@alcatel.de)
Parallel operation of standards A and B, but only one dedicated standard per baseband board. Load sharing based on software reconfiguration is foreseen at the level of the baseband platform.
Load sharing example 1 Load sharing example 2
Wolfgang Koenig is Unit Manager for Advanced Transceiver Solutions within the Alcatel Research & Innovation center, Stuttgart, Germany. (Wolfgang.Koenig@alcatel.de)
Andreas Pascht is Team Leader for Integrated Radio at the Research & Innovation center in Stuttgart, Germany. He is a Distinguished Member of the Alcatel Technical Academy. (Andreas.Pascht@alcatel.de)
References
[1] W. Koenig, G. Luz, S. Walter, U. Weiss: Implementation of a Multi-band Front-end for a Medium Range Base Station within the RMS Project: Test and Measurement Results, SDR Forum Technical Conference, 1618 November 2004, Phoenix, Arizona, USA. [2] T. Bitzer, F. Endress, A. Pascht, M. Schreiter: Imbalance-Compensated Direct Down Conversion Receiver for UMTS Base Stations, IEEE Microwave and Wireless Components Letters, August 2004, volume 14, number 8, page 395-397. [3] S. Walter, M. Bronzel, J. Brakensiek et al: Functional Elements in E2E Reconfigurable Equipment, IST Mobile Summit, June 2004, Lyon, France.
Ulrich Weiss is working on research in the Multi-Band Transceiver Front-End at the Alcatel Research & Innovation center, Stuttgart, Germany. (u.weiss@alcatel.de)
www.alcatel.com/atr
111
TECHNOLOGY WHITE PAPER
ANY ACCESS
F. van der Putten, S. Ooghe
STANDARDIZATION: KEY TO MASS DEPLOYMENT

Standards have evolved to make DSL technology a credible building block for triple-play deployments.
he early 1990s saw the rise of Asymmetric Digital Subscriber Line (ADSL) technology, which proved a giant leap forward in squeezing a higher data rate out of telephone lines compared with the few tens of kbit/s that voice band modems were offering at that time. The believers (with Alcatel in a leading role) put significant efforts into standardizing ADSL as the access technology of the future, based on its potential to bring worldwide information networking to peoples homes. Organizations such as the American National Standards Institute (ANSI T1E1), the Digital AudioVIsual Council (DAVIC) and the DSL Forum all played a part in preparing ADSL to support the High Speed Internet (HSI) service that proved to be the killer application of the late 1990s. Service providers are now starting to deploy triple-play services (data, voice and video). Those relying on DSL access are moving from the traditional voice and HSI services towards interactive video on demand and broadcast TV services. Providers relying on cable access are moving in the same direction, coming from traditional broadcast TV to offer voice and Internet services and interactive video. Eventually, both cable and DSL access providers will compete, offering the same service packages to users. As service providers realize that this scenario is becoming a reality, they are putting increasing pressure on the standardization community to come up with solutions quickly. Over the past few years, the traditional standards bodies and related forums have organized themselves to address key requirements: solution convergence, interoperability, management and promotion. The key players in this process are the International Telecommunication Union Telecommunications (ITU-T), the DSL Forum, the Internet Engineering Task Force (IETF), the Institute of Electrical and Electronics Engineers (IEEE), and the Metro Ethernet Forum (MEF), as well as regional standards organizations such as the Alliance for Telecommunications Industry Solutions (ATIS) in North America, the European Telecommunications Standards Institute (ETSI) in Europe and the China Communications Standard Association (CCSA) specifically for China.
dors, modem vendors, chipset vendors, software providers and solution integrators. Second, and more importantly, it ensures regional and global interoperability of network equipment, which is crucial in making the solution ready for the mass market. And finally, the creation of an open access architecture is essential for the easy adoption of new services from service and content providers.
ITU-T Addresses Fast Changing DSL Technology Requirements

The ITU-T has become the focal point for the development of DSL technology standards, in particular the experts group handling Question 4 in Study Group 15 (Q4/15). It centralizes the efforts of service providers, system vendors and chip makers in a single decision-making process, which, in combination with ITU-Ts improved approval process, shortens the standards time to market. In addition to the traditional need for higher data rates, todays DSL technology standards are being driven by service-related requirements, such as low latency, impulse noise robustness and the transport of Ethernet packets. An overview of recent ITU-T Q4/15 deliverables and ongoing work (summarized in Table 1) is given below. Enhanced SHDSL: Q1/2005 revision of G.991.2 Enhanced Single-pair High Speed Digital Subscriber Line (E-SHDSL) extends SHDSL technology to data rates up to 5.7 Mbit/s over distances up to 1.7 km, making it possible to offer a 10 Mbit/s symmetric Ethernet service over ATM using two copper pairs. As near-end crosstalk occurs in ADSL receivers operating in the same copper binder, E-SHDSL degrades the ADSL performance. To limit any performance degradation, regulatory authorities may choose to limit the E-SHDSL deployment reach.
Why DSL Architecture Standardization?

There are many reasons for deploying a standard DSL architecture. First, it enables broadband evolution to be shaped according to the views and needs of all the players involved, and protects the business interests of operators, system ven112
Near-end crosstalk Unwanted energy transfer measured at one end of a telephone cable between a disturbed pair and another pair where the disturbing signal source is located in the vicinity of the same end.
www.alcatel.com/atr
READSL2: Extending the reach of ADSL technology Reach Extended ADSL2 (READSL2) aims to provide moderate data rates (a few hundred kbit/s) over loops of up to 8 km (0.5 mm wire). This is achieved by transmitting the available signal power over a lower bandwidth. By adapting the used frequency band and power spectrum density to the length of the loop, READSL2 represents a first step in the direction of dynamic spectrum management.
Table 1: Key recommendations and ongoing work in ITU-T
Transceiver specifications
G.991.2 G.992.1 G.992.3 G.992.5 G.993.1 G.vdsl2 G.997.1 Single-pair high-speed digital subscriber line (SHDSL) transceivers Asymmetrical digital subscriber line (ADSL) transceivers Asymmetric digital subscriber line transceivers 2 (ADSL2) Asymmetric Digital Subscriber Line (ADSL) transceivers - Extended bandwidth ADSL2 (ADSL2+) Very high speed digital subscriber line transceivers Very high speed digital subscriber line transceivers 2 (ongoing work) Physical layer management for digital subscriber line (DSL) transceivers
G.998.x ATM, Ethernet and TDM multi-pair bonding ADSL2plus: Hardening G.992.5 against impulse noise ADSL2plus is receiving increasing attention from service providers wishing to offer video services ADSL2(plus) and VDSL, these parameters define boundary from the central office. Several service providers have found that, conditions on minimum/maximum data rate, maximum latency, as they increase DSL data rates, there is increasing sensitivity minimum impulse noise robustness, transmit power and to impulse noise (impairments caused by electromagnetic sigpower spectrum density limitations. nal pick-up in access and in-house wiring). Ongoing ITU-T work focuses on increasing the signal robustness against impulse noise DSL Forum: Enabling Services for all ADSL2plus data rates up to 20 Mbit/s. These will be added over DSL Technologies as implementation options to the ADSL2plus standard and will During its ten years of existence, the DSL be approved in the second quarter of 2005. Forum has played an important role in promoting DSL technology as the enabler for broadband residential and business services, thanks VDSL2: Extending the reach and Spectrum shaping to a combination of marketing and technically-oriented initiarobustness of VDSL The process of tives. The Forums marketing committee covers press contacts The current focus of ITU-T Q4/15 is adapting the Power and showcase participation, with a special focus group for the on the Very high speed Digital SubSpectral Densities European market. The technical committee is organized in a scriber Line 2 (VDSL2) standard. The (PSD) in order to number of working groups, which cover access architecture main drivers for the current work are minimize interference design, home networks, interoperability and network manageextending the reach (up to 2 km) and between neighboring ment. In total, the DSL Forum has published close to a hunimproving robustness against impulse lines. The actual dred technical reports. Figure 1 shows the role played by the noise for high data rate services (25 spectrum shaping DSL Forum in the overall standardization process and the other Mbit/s and above). Unlike the G.993.1 depends on the standards initiatives that relate to the end-to-end broadband VDSL standard, VDSL2 is based only on deployment scenario. access architecture. The key technical reports sourced by the discrete multi-tone technology. As the DSL Forum are summarized in Table 2. deployment of high data rates will require remote deployment, the economics of fiber (FTTx) has an important role, with roll-out plans Architecture design guards end-to-end consistency varying from one operator to another. Spectrum shaping will Migration to services beyond HSI requires the access and be a key feature of the VDSL2 standard to enable aggregation network architectures to evolve accordingly. The mixed deployment of ADSL2plus from the central office design of these new architectures is undertaken by the and VDSL2 from remote cabinets. The Q4/15 experts group Architecture and Transport working group. The framework for evolution to a multi-service architecture aims to complete the VDSL2 standard during the second quaris specified in TR-058, which outlines the service features ter of 2005. required to support the evolution to new applications. It also provides a new service provider interconnection model that Bonding: using multiple pairs for higher data rates allows the access network to be opened up to provide Quality ITU-T finalized three bonding recommendations in the first of Service (QoS) features to third-party application providers. quarter of 2005: ATM, Ethernet and Time Division Multiplex Based on the framework provided in TR-058, a specific (TDM) bonding. Bonding enables multiple copper pairs to be instantiation is provided in TR-059, which describes how new connected to a single user modem, thereby achieving higher services can be offered using the deployed ATM access and data rates. These three bonding standards can be applied to aggregation networks. This is achieved by providing DiffServADSL, SHDSL and VDSL. based IP QoS in the Broadband Remote Access Server (BRAS) and in the DSL modem, called a Routing Gateway. Configuration and management: G.997.1 has the The drivers for further architectural work in the DSL necessary hooks Forum are the migration to Ethernet aggregation and the assoTransceiver configuration and management are defined in ciated support for video services, such as broadcast TV and Recommendation G.997.1, which lists the parameters that can video on demand. When performing interworking between the be configured from outside the transceiver, either by the sysDSL access network and the Ethernet aggregation network, the tem it is built into and/or by the service provider. In
www.alcatel.com/atr
113
DSL Access Multiplexer (DSLAM) becomes involved in the connection establishment process and thus needs to provide several network security features. Furthermore, the DSLAM will be enhanced with multicast features to provide resource-efficient support for video services. This work is currently being progressed in Working Text 101, and is expected to become a technical report that lists the requirements for the DSLAM, Ethernet aggregation network and broadband network gateways.
Figure 1: DSL access architecture standardization and key documents

DSL Forum Operation & Network Management: TR-090 DSL Forum DSL Home: - TR-068 - TR-069 - TR-094 DSL Forum Testing & Interop: - TR-067 - WT-100 - WT-105
IETF IEEE 802.1
IETF IEEE 802.1
IETF IEEE 802.1 DSL Forum Architecture & Transport: TR-092
DSL Modem
Access Node
Aggregation Node
Service Node DSL Forum Architecture
ITU-T IEEE 802.3 IEEE 802.3 & Transport: Home networks enter IEEE 802.3ah - TR-058 the QoS chain MEF - TR-059 The increasing importance of QoS - WT-101 DSL Forum Marketing for delivering triple-play services has raised the interest in routing gateways and home networks. A framework for the delivery of multiple services to the home is functional testing in TR-048 and its update TR-067. Recently, described in TR-094. To support new services in the home, a European Customer Premises Equipment (CPE) logo protodays simple DSL bridged modem will evolve to become routgram has been defined, based on meeting the TR-067 Euroing gateways, the basic requirements for which are listed in pean requirements. Starting early in 2005, the main focus of TR-068. Additionally, the triple-play service offering requires work is shifting towards interoperability testing for ADSL2plus configuration and performance monitoring of the routing gatein WT-100 and towards a more detailed functional test suite way and home terminals. The DSL Forum has issued TR-069, documented in WT-105. defining an eXtensible Markup Language (XML) based mechanism for managing the routing gateway and home terminals IETF from an auto-configuration server. Currently, special attenThe IETF is concerned with evolution of tion is being paid to the management of Session Initiation Prothe Internet architecture as a whole. With HSI tocol (SIP) phones in the home, as an example of home appliaccess being the first killer application for ance management. ADSL, the broadband access architecture traditionally relies on a number of key Request For Comments documents (RFC), such as the IP connectivity establishment protoInteroperability to foster the mass market cols (Point-to-Point Protocol, PPP; PPP over ATM, PPPoA; PPP The DSL Forum Test and Interoperability Working Group over Ethernet, PPPoE; Dynamic Host Configuration Protocol, is setting the requirements for DSL interoperability. This DHCP; and Layer 2 Tunneling Protocol, L2TP) and Authenticamainly consists of data rate performance under various loop tion, Authorization and Accounting (AAA) procedures (Remote and noise conditions, together with some limited ADSL Access Dial-In User Service; RADIUS). Table 2: Key technical reports and ongoing work in the DSL Forum With the evolution to triple-play access architectures, the work perMarket requirements formed by the IETF is becoming even more relevant to the access and aggreTR-58 Multi-service Architecture and Framework Requirements gation network equipment. Specifically Architectural requirements the DSLAMs, BRASs and IP routers will need to support a number of new protoTR-59 DSL Evolution Architecture Requirements for Support of QoS enabled IP Services col procedures for voice, video and busiTR-94 Multi-Service Delivery Framework for Home Networks ness services. These include procedures for multicast support, new authentication WT-101 Migration to Ethernet Based DSL Aggregation schemes and support for IP virtual priEquipment specifications vate networks. Moreover, the IETF has worked out a new IP addressing TR-67 ADSL Interoperability Test Plan (Formerly TR-048) architecture known as IPv6, which TR-68 Base Requirements for an ADSL modem with routing will become of increasing importance as more and more IP TR-69 CPE WAN Management Protocol devices are connected. It is expected TR-92 Broadband Remote Access Server Requirements that the work performed in many IETF working groups will find its way into the WT-100 ADSL2plus Interoperability Test Plan future DSL access architecture being WT-104 Provisioning Parameters for VoIP CPE defined by the DSL Forum Architecture and Transport working group. WT-105 ADSL2plus Functional Test Plan
114
www.alcatel.com/atr
IEEE
IEEE is a leading authority in many technical areas, including computer engineering and telecommunications. The IEEE 802 Working Group focuses on aspects ranging from wired and wireless Ethernet transmission technologies (802.3 and 802.11/802.16, respectively) and resilient packet rings (802.17) to higher layer Local Area Network (LAN) protocols, such as bridging and switching (802.1). In the case of DSL-based access architectures, Working Groups 802.1 and 802.3 provide a number of crucial standards: The IEEE 802.3 Working Group develops Ethernet standards. It has standardized a number of 1 and 10 Gigabit Ethernet transmission interfaces that support bandwidth intensive video services. Recently the group adopted 802.3ah, also known as Ethernet in the First Mile, specifying the transport of Ethernet frames over broadband access technologies, including SHDSL, VDSL and Passive Optical Networks (PON). The IEEE 802.1 Working Group develops standards and recommended practices for local and metropolitan area networks, link security and network management. These include IEEE standards 802.1D and 802.1Q, allowing Ethernet bridging with and without support for Virtual LANs (VLAN). To provide a higher level of scalability, 802.1ad extends the principles of 802.1Q by specifying procedures to add two VLANs to Ethernet frames. To make Ethernet a carrier-grade technology, 802.1ag is specifying connectivity fault management features, which make it possible to perform Operations, Administration and Maintenance (OAM) functions similar as those currently used in ATM networks. Finally, 802.1X provides procedures for performing portbased access control in an Ethernet network; data confidentiality, frame data integrity and data origin authenticity are possible using 802.1AE.
(IMS), which describes the overall architecture, is based on the Third Generation Partnership Project (3GPP) core IMS. The IMS architecture includes a network attachment subsystem, which allows authentication, user profile based authorization and access network configuration, IP address allocation and location management. Similarly, the resource and admission control subsystem is part of the IMS architecture; it allows dynamic resource admission control and bandwidth reservations by controlling packet filters in the data plane. Using these components, the TISPAN NGN architecture paves the way for convergence between fixed and mobile networks. Release 1 of the TISPAN NGN is expected to be approved in 2005.
Conclusion
Over the past ten years, the access network has evolved in the three dimensions that are fundamental to the successful offering of services: high data speed, high data quality and high data sustainability. In standardization terms, these are the data, control and management planes, the evolving requirements of which have been (and still are) at the heart of a changing standardization framework. The data speed has evolved from voice band modem speeds (a few tens of kilobits per second) to multi-megabit per second speeds. This has required access technology standards to initiate different tracks for ADSL (residential focus), SHDSL (business focus) and VDSL. The last of these is currently considered as the ultimate speed technology, offering tens of megabits per second, but it will require deeper fiber penetration in the access network. Data quality has evolved from best-effort Internet access services to real-time services requiring QoS control in the access network. Consequently, the access equipment (especially the DSLAM) must become Ethernet and IP aware. This emphasizes the growing importance of related access standards. Data sustainability, which used to consist of DSL management and hence mainly DSL-focused troubleshooting, has broadened its scope to include service configuration and performance monitoring across several network elements, including routing gateways and home network appliances. It has become apparent that the evolution to triple-play services involves more than a migration to higher data rates. Thus the standardization framework has had to be broadened, and still needs to fill in building blocks for mass market endto-end service delivery. The ITU-T, DSL Forum, IETF, IEEE and MEF are all engaging closely in this evolution.
Metro Ethernet Forum

The MEF is a relatively new organization. Founded in June 2001, its initial charter was to promote the adoption of Ethernet technology, focusing on Ethernet access for business customers. Its current mission is to accelerate the adoption of optical Ethernet as the technology of choice in metropolitan networks worldwide. The objectives of the MEF are to define Ethernet services delivered over a metropolitan network, and to define carrierclass Ethernet transport technologies by specifying the architecture, protocols, management and OAM required to support Ethernet services in Ethernet metropolitan networks. Similar to the DSL Forum, the MEF is organized into a marketing and a technical committee. In its four years of existence, MEF has created over ten technical specifications, providing a general Ethernet architecture and focusing on Ethernet services, traffic management, protection and circuit emulation. Some of the carrier-grade features that are currently being worked on are Ethernet service OAM and the specification of Ethernet UserNetwork Interface (UNI) enhancements, such as the Ethernet Local Management Interface (LMI).
ETSI TISPAN
The TISPAN Next Generation Network (NGN) project is developing the specification for an architecture that supports voice services over packet networks, such as DSL. The TISPAN IP Multimedia Subsystem
Frank van der Putten is Corporate Technical Standards Manager for Wired Access Networks and Director Standardization for the Alcatel Access Network Division, Antwerp, Belgium. He is a Fellow of the Alcatel Technical Academy. (Frank.van_der_putten@alcatel.be)
Sven Ooghe is responsible for access functional strategy and coordinating broadband architecture standardization works in the Chief Technology Office team, Access Network Division, in the Alcatel Fixed Communication Group, Antwerp, Belgium. (Sven.ooghe@alcatel.be)
www.alcatel.com/atr
115
ANY ACCESS
Th. Pfeiffer, E. Ringoot, A. Granger, D. Wang
OPTICAL FIBERS PAVE THE WAY TO FASTER BROADBAND ACCESS

Emerging services and increasing competition are forcing carriers to deploy optical fibers in the access network, ultimately serving each subscriber over a dedicated fiber link.
meters (Very high speed Digital Subscriber Line; VDSL). Finally, here is an increasing demand for high bandwidth connecfibers may have a role in backhauling for mobile and fixed wiretions in markets around the world. However, upgrading the less networks. bandwidth of traditional technologies, like Digital SubTo optimize fiber use, point-to-multipoint networks are genscriber Line (DSL), coaxial cable and wireless connections, is erally used with a few fibers running from the central office to becoming ever more challenging as these technologies get closer to their practical capacity limits. Competition between network Figure 1: Basic fiber access architectures operators is compelling them to differentiate their offers by providing more PON Optical Ethernet Optical Ethernet and better services to their customers, passive star active star point-to-point which in turn requires the latest future-proof technologies. OLT ONTs The delivery of triple-play services (data, voice and video) at competitive prices is essential in this market enviRemote Central Office Central Office Central Office Node ronment. Current fiber access deployments in Japan and the USA offer subscribers broadband access at prices in the range of two to three times DSL service prices, nodes in the neighborhood from where multiple drop fibers prodepending on the bitrate and market. vide the connections to users or to VDSL nodes. In a Passive Another factor is politics, which is playOptical Network (PON), as shown in Figure 1, signals from the ing an important role in steering broadOptical Line Termination (OLT) at the central office are optiband deployment. Favorable regulacally broadcast via the feeder fibers into the drop fibers to the tions, like the decision of the US FedOptical Network Termination (ONT). Upstream, Time Division eral Communications Commission Multiple Access (TDMA) based protocols ensure collision-free (FCC) not to require unbundling of transmission from the ONT to the OLT. Alternatively, the active fiber lines in the access network, star configuration employs Ethernet switches at the remote sound national broadband policies and nodes to distribute and aggregate the traffic to and from users. the launching of national programs or Here, the layer 2 protocol is based on Ethernet. As a third funding of broadband deployments, are option, in point-to-point networks single fiber links connect helping in this respect. each subscriber directly to the Marketing is also an important central office. factor; novel communication technologies are increasingly viewed as a means of increasing the attractiveness and accelFiber to the x Today erating the economic development of municipalities, regions Existing FTTx solutions and even countries. One existing solution is the Broadband PON (BPON), an Increasing the bandwidth to users means bringing Asynchronous transfer mode fibers closer to the subscribers. In addition to the ultimate PON (APON) enhanced by an solution of Fiber to the Home (FTTH) sometimes also called optical overlay channel for video Fiber to the Premises (FTTP) or Fiber to the User (FTTU) services. This widely deployed there is considerable potential for early cost-effective broadand mature technology is band access deployments using hybrid solutions that combine currently supported by the fiber for the feeder section (Fiber to the Node, FTTN; an active Alcatel 7340 FTTU. node in the neighborhood) and copper for the final tens of
116
www.alcatel.com/atr
Table 1: PON comparison

ITU-T BPON Data rate (Mbit/s) Line coding Minimum split (on TC layer) Maximum split (on TC layer) Maximum logical reach supported by TC layer Layer 2 protocols Standards documents TDM support Typical downstream capacity (for IP data throughput) Typical upstream capacity (for IP data throughput) OAM Downstream security AES: Advanced Encryption Standard OAM: Operations, Administration and Maintenance down: 1244, 622, 155 up: 622, 155 NRZ (+ scrambling) 32 64 20 km ATM ITU-T G.983 series TDM over ATM 520 Mbit/s (for 622 Mbit/s line rate) 500 Mbit/s (for 622 Mbit/s line rate) PLOAM + OMCI Churning or AES ITU-T GPON down: 2488, 1244 up: 2488, 1244, 622, 155 NRZ (+ scrambling) 64 128 60 km (with 20 km differential between ONTs) Ethernet, TDM over GEM (GPON Encapsulation Mode), ATM ITU-T G.984 series native TDM, TDM over ATM, TDM over Packet 1170 Mbit/s (for 1.244 Gbit/s line rate) 1160 Mbit/s (for 1.244 Gbit/s line rate) PLOAM + OMCI AES (counter mode) IEEE EPON down: 1250 up: 1250 8b/10b 16 not specified 10 km, 20 km Ethernet IEEE 802.3ah TDM over Packet 910 Mbit/s 760-860 Mbit/s Ethernet OAM (+ optional SNMP) not defined SNMP: Simple Network Management Protocol TC: Transmission Convergence
OMCI: ONT Management and Control Interface PLOAM: Physical Layer OAM
It carries analog and interactive digital TV together with voice and data services. Today BPONs are typically deployed with 622 Mbit/s in the downstream direction and 155 Mbit/s upstream, giving each user a downstream capacity of 20 to 30 Mbit/s. The International Telecommunication Union Telecommunications (ITU-T) BPON standard, with its higherlayer functionality, including security, management, flexible configuration and Quality of Service (QoS), ensures straightforward interoperability. As a result of its maturity and wide deployment, BPON costs have dramatically decreased, making it an attractive choice for service providers contemplating deploying a fiber access solution. Media Another is Ethernet point-to-point Converter (MC) fiber, including the Japanese media Optical point-toconverter and IEEE 802.3ah (Ethernet point transmission in the First Mile; EFM) optical pointtechnology for to-point technology. Media converter access networks. technology, which was standardized in 2002 by the Telecommunications Technology Committee (TTC) as TS-1000, has been widely deployed in Japan. In contrast, EFM only recently released optical point-to-point physical layer specifications for access networks (symmetrical 100 Mbit/s and 1 Gbit/s). Next generation FTTx solutions Ethernet PON (EPON), which was ratified in September 2004 as part of IEEE 802.3ah, supports symmetric Gigabit Ethernet and is compatible with existing Ethernet networks. (In Japan, the term EPON refers to a proprietary 100 Mbit/s PON, whereas the IEEE EPON is called GE-PON.) One of the main concerns with EPON is that the standard does not specify all the features needed to transport carrier-grade services, so proprietary solutions are needed to make it compatible. Moreover,
it offers no bitrate flexibility on the physical layer and the bandwidth efficiency is low because of inefficient line coding and large overhead. However, simple requirements regarding the physical components and service features favored the early availability of commercial EPON components. Gigabit PON (GPON) was standardized by ITU-T in June 2004. The GPON standard includes a variety of line rates up to 2.488 Gbit/s, symmetric and asymmetric. Together Guard time with much reduced coding loss and In Time Division shorter guard times (i.e. time slots Multiple Access between adjacent optical packets from (TDMA) networks, different ONTs), the net bandwidth of like PONs the GPON is much higher than in (upstream EPON. Besides transporting native direction), optical Ethernet data, the GPON also effipackets from ciently transports multimode services different subscri(TDM, voice and ATM). Today, GPON bers are sent to the deployment is hampered by the late central node in availability of the basic component such way that they technology. However, it is set to catch arrive sequentially up with the early EPON deployments. at the central Table 1 summarizes the main charreceiver. acteristics of these PONs. Successive packets are separated from each other by dark Worldwide market periods (guard North America times) several 10 Two operators started deploying 100 bits long. FTTP in 2004 using BPON technology. During this time the Verizon announced its vendor selection receiver is reset to in November 2003 and started deployadapt to receiving ment in 2004. Similarly, SBC launched a packet from the its Lightspeed project in late 2004, next subscriber. with Alcatel as the lead supplier (see the Editorial in this issue). It has
117
www.alcatel.com/atr
announced plans for 18 million households passed by 2007 using FTTP and FTTN + VDSL. Shortly after, BellSouth granted Alcatel an award for the delivery of FTTN equipment. East Asia Japan is the most established FTTH market with more than 1.6 million subscribers by August 2004. APON, BPON and (proprietary) 100 Mbit/s EPON are the technologies being deployed. Although the Japanese standard Media Converter has been installed in the past, it is losing ground to xPON technology. In Korea, 8.5 million subscribers will have access to Fiber to the Cabinet (FTTC) + VDSL networks in 2005. Korea Telecom is now starting to move to FTTH, using Wavelength
Future PON Evolution

BPON with its 20 to 30 Mbit/s throughput per subscriber is a future-proof fiber access solution, although the availability of Gbit/s PONs will lead to a transition in the deployed technology. Since innovation never ceases, the search is on for further enhancements leading to the next-but-one generation of PONs. Next-but-one generation PONs Although commercial GPON and EPON systems are still under development, upgrades are being studied to increase the bitrate, link lengths and splitting ratios. This will help to reduce operators costs by concentrating and reducing the quantity of central office equipment and fiber cabling needed in the feeder sections of their networks. At the physical layer, current PON standards include WDM to separate the upstream (1310 nm) and downstream (1490 nm) and to provide an additional overlay channel (1555 nm e.g. for video). WDM could, however, also be used to enhance the capacity for single users by providing a dedicated high-speed optical point-to-point channel and/or to realize several GPON networks over a single feeder fiber using different pairs of wavelengths for the different GPONs (see Figure 2). Many configurations can be conceived to increase network capacity and topology diversity if wavelength is included as a design parameter. Other technological options are being considered (e.g. optical repeaters) to increase the reach of the PON to around 100 km or to increase the splitting ratio to 1:1000, thus paving the way to reducing the number of central offices. These advanced approaches need further research before they can be standardized. Near-term research and development topics relate to PON improvements in terms of capacity optimization and cost reduction. PONs will not always look as simple as above, but in many cases will be composed of groups of subscribers located near to the OLT after just one splitter stage, while others are further away behind two or more cascaded splitters (see Figure 3). The capacity of the entire network can be max-
Figure 2: Wavelength enhanced PON
WDM Splitter
1 / 5 2 / 6 3 / 7
1 4 5 8
OLT
4 / 8
Wavelength allocation (CWDM) downstream upstream
1 2 3 4
1471 nm
5 6 7 8
1611 nm
CWDM: Coarse Wavelength Division Multiplexing
Division Multiplexing (WDM) to provide 50 to 100 Mbit/s to each household. A number of Chinese carriers have begun EPON trials, whereas FTTC + LAN (or + VDSL) is regarded as more appealing by competitive local exchange carriers. Together with India, China will become the worlds largest FTTx market in the near future. Rest of the World FastWeb in Italy has more than 400 000 broadband subscribers in six cities, with some 180 000 served by direct fiber connections. Sweden also has several small FTTH networks; the biggest Bredbandsbolaget (B2) has about 120 000 subscribers. Services include Voice over Internet Protocol (VoIP), data and streaming video. In northern Germany, EWE started to deploy FTTH in 2004. EnergiMidt in Denmark also has significant PON deployment plans. As part of its 21CN initiative, British Telecom has started FTTP field trials for 3000 residential users and more than 100 business users. Telstra (Australia) will soon launch a commercial FTTP pilot trial in Queensland using the Alcatel 7340 FTTU (BPON) solution.
118
Figure 3: Bus topology for the PON

asymmetric coupler 25% 1:8 1:8 asymmetric coupler 33%
OLT
ONTs
www.alcatel.com/atr
Figure 4: FTTU PON with CATV video overlay transport imized if the splitting ratios are adapted accordingly, thus equalizing the optical power budget for all subCentral Office Passive Optical Home Network Network (BPON) scriber groups. Alternatively, ONTs with better power budgets could be to OSS assigned higher line rates than others. Voice 1:32 split This can be implemented by providing HSI splitters with unbalanced or even tunOLT IP WDM able splitting ratios. However, multiCATV CATV 1490 nm Network ONT over coax DBS bitrate or otherwise flexible PON operSTB ation is not yet incorporated in the Voice 1310 nm 1555 Network standards. nm Finally, it should be mentioned that the optical components, especially the optical transceiver modV-OLT Video ules, are still the most expensive sinHead-end gle part of the network. Long-reach BPON: Broadband Passive Optical Network STB: Set-Top Box networks with bitrates in the Gbit/s OSS: Operations Support System V-OLT: Video Optical Line Termination, range cannot be operated using costefficient Fabry-Perot lasers because of dispersion-induced noise problems. This might be mitigated either by introducing low cost Verticommon IP infrastructure, allowing operators to deliver cal Cavity Surface Emitting Laser (VCSEL) technology, or by new user-centric multimedia experiences that provide applying Forward Error Correction (FEC) to reduce the bit additional revenue-generating opportunities. error ratio in the electronic domain after reception. This is currently being investigated to evaluate whether it can lead to more CATV DBS PON overlay cost-effective solutions. CATV video channels are carried over the third PON wavelength range around 1550 nm, similarly to CATV HFC networks (see Figure 4). To provide the signal to a number of PONs, an Video Service Delivery over FTTU optical amplifier system (video OLT) boosts the video optical Triple play has become the universal business model for signal before merging it, via an optical multiplexer (WDM), with telecommunication operations; the video services compothe downstream data PON signal at 1490 nm. The ONT video nent is decisive for the operators business performance, receiver extracts the video channels and sends them over the regardless of the type of broadband support network. users coaxial home network. The intrinsic unlimited bandwidth of fiber optics makes As the video signal is optical end-to-end, it is much less FTTU a suitable networking framework for developing triple subject to distortion and noise than in HFC, leading to a higher play and, more particularly, video services as it overcomes the quality service. On average, the channel capacity is 80 anabandwidth limitations and legacy constraints of copper, coaxlog channels and 200 MPEG-2 streams, typically supported ial and wireless networks. by 33 digital radio frequency channels employing 256 QuadFTTU allows the direct emulation of Hybrid Fiber Coax rature Amplitude Modulation (QAM) (USA). The system is (HFC) Cable TV (CATV) broadcast services, together with the open to operating CATV analog and digital worldwide standelivery of voice and High Speed Internet (HSI) via a single dards and to the migration to all digital QAM or Orthogonal fiber all the way to the home. In addition, it supports the conFrequency Division Multiplex (OFDM) channels with aggrevergence of video, voice and Internet delivery over a gate data rates up to 5 Gbit/s. Similar solutions apply to the delivFigure 5: FFTTU PON with in-band video transport ery of Direct Broadcast Satellite (DBS) content to the home, thus avoiding the hassle of dish installations and layout.
Central Office Passive Optical Network (GPON) Home Network
to OSS 1:32 split IP Network for his video voice service OLT 1490 nm 1310 nm ONT 100/BT Over Cat 5 or MOCA CATV DBS STB Voice HSI
HSI: High-Speed Internet
MOCA: Multimedia Over Coax Alliance
Overlay versus in-band The key advantage of CATV overlay is that the existing standard analog customer premises equipment can be reused. Because of the rapid development of digital video, a set-top box is generally required, paving the way to alternative solutions, such as IP TV, via in-band transport over the PON (see Figure 5). Compared with the standard fixed broadcast bundle of CATV, in-band allows a seamless connection to Web or dedicated video servers which are accessible via the IP network.
119
www.alcatel.com/atr
Again, an FTTU platform is an ideal solution to carry in-band IP Video traffic in greenfield situations and to achieve convergence with IP data and voice. Migration to GPON provides a large transport capacity.
Figure 6: Redundant PON configurations: full redundancy (Type C, top); OLT and root redundancy (Type B, bottom)
ONU #1 NT (1)
Network Availability
NT (0)
N:1
Redundancy and protection switching Availability of the last mile connecN:1 ONU #N tions1 will play an increasingly imporNT (1) redundant edundant d d t tant role in future fiber-based access optical networks as they provide paying (e.g. NT (0) splitters TV) or mission critical (e.g. security systems) services to a large number of users. This has been taken into account redundant redundant redundant in the early ITU-T standardization of NTs feeder drop PONs. Different protection architectures have been specified in the BPON ONU #1 standard, based on principles similar to those for high-speed core networks. NT (1) The most secure, but also most NT (0) expensive, way to protect the network from failures is by duplicating each N:2 optical transceiver, each splitter and ONU #N each fiber section (type C in Figure 6). optical ti l i splitter The various possible paths between the NT (1) OLT and ONTs may be configured in NT (0) either a 1:1 or 1+1 protection mode. Special protocol precautions ensure correct switchover between different unprotected redundant PON sections when a failure is NT and drop feeder detected. This expensive redundant network architecture may be needed in LT: Line Termination NT: Network Termination cases where the PON provides connectivity to high-end users with high availability requirements or to nodes that are shared by a large numThe conventional technology for ber of users, as found in large multiple dwelling units. fiber diagnosis is Optical Time Domain For the residential and Small and Medium size Enterprise Reflectometry (OTDR). It senses the (SME) markets, a reduced redundancy (if any) that protects fiber attenuation by injecting a short the feeder section of the PON, but not the final drop section, optical test pulse and measuring the would be sufficient (type B in Figure 6). This is similar to clasoptical power backscattered from the sical phone networks in which the final drop section from the fiber. This technique provides informaexchange to the subscribers premises is an unprotected tion about the amount and location of point-to-point link. attenuation or reflection along the fiber length (see Figure 7). Conventional OTDR equipment is, however, Physical Link Monitoring in Access expensive and is generally not compatTo initiate protection switching or proactive mainteible with an operating PON system. nance actions, the network management system needs detailed information about the link status. Today this inforAlcatel is developing solutions mation is derived from global status messages for the line equipfor integrating OTDR functionality ment (e.g. received signal power). To anticipate potential probinto the optical line equipment of lems and to better identify the causes of problems, network a PON and making it an integral operators need to look deeper into the physical network infrapart of the network and managestructure and continuously monitor the performance of the ment system. This enables OTDR fiber links. The resulting enhanced reliability of network measurements to be performed operation benefits the operators business by increasing remotely, continuously and at low cost service stability, thereby avoiding the need for rebates when from every OLT and ONT in the netservice-level-agreements are breached. work. PON-specific OTDR problems, like reduced sensitivity and ambiguous Last mile and first mile refer to the same network segment; the former traces caused by superposition of originates from the telecommunication industry, and the latter from the datacommunication industry. backscattered powers from the differ1
redundant LT
OLT
redundant LT
Backscattered power State-ofthe-art optical fibers attenuate the injected light signals mainly by Rayleigh scattering which is caused by microscopic variations in the refractive index of the fiber along its length. Part of the scattered power is collected by the fiber itself in the backward direction and can be used to analyze the attenuation characteristics from the input end of the fiber.
120
www.alcatel.com/atr
ent drop fibers in a PON, are eliminated. Future WDM-based PONs can be easily monitored without the need for wavelength tunable test equipment, since the measurements will be performed by the WDM compliant optical data transceivers.
Figure 7: Optical data transceiver modules in OLT will include (left) OTDR functionality; (right) a typical OTDR trace of a 10 km point-to-point fiber link
Open connector Fiber junction (diff. backscatter coefficient)
Conclusion
Many operators are now increasing their investment in fiber access networks, driven by higher service bandwidth requirements as well as by increased competition and political (de)regulation. Early deployments have started in North America and East Asia, and will shortly be followed by deployments in other countries and regions. National broadband policies are helpful in accelerating this evolution. However, they should not focus solely on general broadband access, but more specifically on the need for fiber optic networks. Fiber technology will be a differentiator for operators, for cities and even for countries, delivering benefits to citizens and network operators alike. Migration of large networks to fiber needs to be implemented in a well focused and coordinated fashion in order to reach a large number of subscribers in a short time frame. Fiber technology should be strongly pushed for network installation in new-build areas now that deployment costs are similar to those of traditional solutions. For existing builds, the speed of introducing FTTP will be slower, and will depend on the existing outside plant (buried, aerial) and on population density (suburban, urban, etc). Technology-wise, many innovative solutions for further enhancements and cost reductions of fiber access networks are on their way to realization, and are being studied in research labs worldwide, including those at Alcatel. 100 s (10 km fiber)
Thomas Pfeiffer is Project Leader, Optical Access, and a Distinguished Member of the Alcatel Technical Academy, located at Research & Innovation, Stuttgart, Germany. (Thomas.Pfeiffer@alcatel.de) Edwin Ringoot is Product Manager, and a Regular Member of the Alcatel Technical Academy, in the Alcatel Access Networks Division, Antwerp, Belgium. (Edwin.Ringoot@alcatel.be)
Alain Granger is Principal Engineer and a Distinguished Member of the Alcatel Technical Academy, in the R&D Competences & Processes Optics department of the Access Networks Division in Raleigh, North Carolina, USA. (Alain.Granger@alcatel.com) Dai Wang is Strategy Manager, Broadband Access, in the Network Strategy Group, Strategy Development Department, Shanghai, China. (Dai.Wang@alcatel-sbell.com.cn)
www.alcatel.com/atr
121
ANY ACCESS
A. Bertout, J. Couet
EXTENDING BROADBAND REACH BY SATELLITE

Combining broadband satellite solutions with terrestrial wireline and wireless solutions can rapidly extend the reach of IP multimedia applications to underserved parts of the world.
which users can enjoy services and interact seamlessly over n many regions of the world, the cost of broadall networks using a variety of terminals. band services, and especially data communication, is too high for even small businesses to afford. Telecom services prices in such regions are significantly Satellite-based Architectures: higher than in more developed countries, thus significantly Universal Broadband Access Solutions handicapping the expansion of services to users worldwide Services to fixed and on the move users and putting a break on the global economy. Nevertheless, Generalized access to broadband services is of prime there is a strong demand for a variety of broadband servimportance in the move towards a user-centric broadband ices from individual users (consumers and professionals) world; this means that services must be dependable, available and groups of users. almost everywhere and attractively priced. Whether they are local inhabitants or only roaming in these For a long time, satellites have been successfully delivering regions, users deserve the same services and to be able to video services both to broadcast local loops and directly to peointerface with their services in the same way irrespective of ples homes. However, although they are reliable and can be their location. This combination of identical perception, deployed rapidly, in the case of voice and high speed data servubiquitous access and a common interface is also known as ices satellites are still perceived as costly solutions reserved user-centric broadband. In this respect, wireless technologies, for niche markets. Nevertheless, technology has now signifiand satellite technologies in particular, are well positioned to cantly evolved, enabling satellites to play an important part in facilitate the rapid deployment of a broadband infrastructure the infrastructure used to deliver not only video, but also other at a relatively low initial cost. To help reduce this digiuser-centric broadband services. The benefits brought by Digital Video Broadcast tal divide, governments and regulators are increasReturn Channel via Satellite (DVB-RCS), the first true staningly prepared to free up satellite frequency specdard for high data rate unicast services, are at the heart of trum for broadband services. this assertion: The list of user applications encompasses basic fixed and mobile voice and data services, and can be extended to applications like credit Figure 1: Connecting isolated users to public broadband networks card authorization, access to corporate networks and distance learning. FurUser Density thermore, the inherent broadcast capabilities of satellite solutions should be leveraged to develop personalized content delivery to users and/or to regional Direct Isolated Connections servers connected to the terrestrial network infrastructure. Even though the cost of satellite bandwidth is sometimes seen as a limBackhaul iting factor in the deployment of Connections individual satellite-based services, broadcast services benefit from Scattered unmatched reach using a single transmission. This can swing the overall Any Wireless, WiMAX, Broadband cost of services in favor of satellite, or PLT, DSL,.. Network especially when it comes to tripleWiFi play (i.e. voice, video and data) servGrouped ices. This is a significant step towards a user-centric broadband world in
122
www.alcatel.com/atr
Figure 2: Integrating remote enterprise users into existing broadband private networks Being based on the Digital Video Broadcast Satellite (DVB-S) standard for the downlink (network to terminals), which is extensively used today by Digital Broadcast Satellite Operators (DBSO), DVB-RCS technology has the advantage that it can easily mix broadcast, multicast and two-way traffic, enabling attractive service mixes to be delivered to users. For the uplink, DVB-RCS provides VPN A efficient transmission and achieves excellent carrier loading factors based on contention-free advanced resource allocation mechanisms adapted to various types of traffic VPN B (up to 4 Mbit/s). Full compatibility with all Internet Enterprise Users Enterprise Users at central Protocol (IP) based applications. at remote sites Approved in 2000 by the European sites Telecommunications Standards InstiVPN: Virtual Private Network tute (ETSI), DVB-RCS is now a fully implemented and tested multi-vendor technology; In its new release, which was finalized in December 2004 Figures 1 and 2 show examples of this combination/inteand supported by Alcatel Space, a key participant in DVB gration to provide access to existing services delivered by pubForum working groups, the standard proposes DVB-S2 for lic broadband networks and to integrate remote sites into existthe downlink. In the broadcast mode, ing private networks. DVB-S2 realizes a spectral efficiency Unicast / gain of about 35% compared with Alcatel 9780 DVB-RCS Multicast / DVB-S. DBSOs will be able to capital(Digital Video Broadcasting Return Channel via Satellite) Broadcast ize on this significant improvement to Generic Alcatel deliver high definition TV. However, transmission modes the key advantage comes from the involving point-tointeractive mode implementation point or point-tothat allows real-time adaptive coding multipoint and modulation according to the communications propagation conditions at each indibetween parties. vidual terminal. Such user-centric Broadcast networks will no longer be sized on a One-to-all worst case basis, leading to signifitransmission where cant savings in satellite bandwidth as the source sends the satellite transponder throughone copy of the put is more than doubled compared content to all with DVB-S. nodes, whether Alcatel considers DVB-S2/DVBthey wish to receive Alcatel Space AMC12 Broadband Satellite RCS to be a key technology it or not. Alcatel Space - Photo Basile which will power its satelliteMulticast based solutions for fixed broadTransmission from band services (Alcatel 9780 DVBone source to a RCS). Nevertheless, these solutions group of recipients require more than pure satellite techvia a single nology to contribute to the user-centransmission tric broadband infrastructure. Smooth resource. integration with core terrestrial netUnicast works, flexible network isolation and Point-to-point easy combination with terrestrial local transmission loops, for example, WiFi, WiMAX and requiring the DSL (via micro DSL access multiplexsource to send an ers), are important advantages and individual copy of help to provide the technology mix the content to each required to meet users needs for requester. broadband connectivity.
www.alcatel.com/atr
123
End-to-end satellite systems for on the move users over mobile networks Broadband satellite access technologies can also be used to serve mobile vehicles in which many users could access their mail, telephone or other broadband services. The MObile Wideband GLobal Link sYstem (MOWGLY) Sixth Framework Program European project, led by Alcatel Space, is studying the development of and complements to the DVB-S2/DVBRCS broadband technologies that are needed to connect users while traveling in planes, trains and cruise ships.
Figure 3: End-to-end satellite systems for users traveling in mobile vehicles
DVB-S2 ISP Servers LAN NOC DVB-RCS OSS Remote GMSS ISP DVB-RCS Gateway Ka Band Spots Ku Band Spots Interconnection Backbone ISP Internet
>
www.mowgly.org: MObile Wideband GLobal Link sYstem (MOWGLY).

LAN LAN
Internet
The project will test a number of solutions, including satellite terminals and antennas tailored to the various mobile environments. In particular, an advanced medium access control technique for a group of passengers using a shared satellite terminal will be developed, together with the
Optional terrestrial return link
GMSS: Gateway Management SubSystem ISP: Internet Service Provider LAN: Local Area Network
NOC: Network Operations Center OSS: Operations Support System
Mobile Wideband Global Link sYstem (MOWGLY)

The MOWGLY project is studying the implementation of new standards for the provision of broadband satellite access to users in aircraft, trains and ships.
necessary Quality of Service (QoS) control and networking technologies (Figure 3). Bringing together leading actors in the field of user transport, this project will focus on user needs and the required performance. User groups expect to be able to use their standard mobile handsets and portable PCs, as well as their existing subscriptions and service preferences, while on board. 2G /3G mobile services and beyond 2G/2.5G backhaul by satellite The transport of new 2G/2.5G mobile services is shifting from circuitswitched to packet-switched data. IP has become the common denominator for transporting all fixed and mobile services. Broadband IP access by satellite is a promising backhaul technique for the inexpensive connection of 2G/2.5G mobile base stations to the core network. Service areas can be quickly extended to include distant cities, where many fixed and mobile users can be offered Global System for Mobile communication (GSM) and other mobile services via satellite very effectively. Optimization of the satellite spectrum using Demand Assignment Multiple Access (DAMA) technology enables 2G/2.5G mobile traffic to be
Figure 4: 2G/2.5G GSM backhaul by satellite
Remote Stations Multimode User Terminal UMA / WLAN DVB-RCS Compressor Terminal IP Compact GSM BTS DVB-RCS Compressor Terminal DAMA + Adaptive waveform
Shared Hub DVB-RCS Hub DAMA + Compressor IP IP BSC Ater UMA UNC Ater
IP or E1
E1
BSC
Ater
GSM BTS DVB-RCS Compressor Terminal
Drop-Insert daisy chain
BSC: Base Station Controller BTS: Base Transceiver Station UNC: UMA Network Controller
124
www.alcatel.com/atr
three million users and Sirius Radio more than one million users. The expectation is that by 2010 SDR sysSMALL LOW EARTH ORBIT (LEO) SATELLITES BIG GEO SATELLITES tems will serve about 30 million users High power Low power Large antenna Small antenna in North America (Lehman Bros, 2004). Regional coverage Worldwide coverage These SDR systems are based on a Low deployment cost High deployment cost hybrid architecture combining satellite - 2G systems: 12 m - 3G projects: 22m broadcasting with, especially in urban - B3G future: > 30 m zones, terrestrial repeaters to ensure seamless reception for vehicles when Associated the satellite is not within the line of with terrestrial sight. XM Radio uses about 700 terresrepeaters trial repeaters deployed in the 60 largest US cities. Consequently, SDR broadcast systems put vehicles at the Users forced to use Users can use center of an hybrid open communicasatellite-specific terminal standard size terminals tion environment, including terrestrial and satellite broadcast components and a point-to-point terrestrial GSM / General Packet Radio Service (GPRS) / 3G component. distributed efficiently via satellite. The new DVB-RCS standard This integrated SDR system comprises three layers (see implements DAMA for high-speed Internet access services (see Figure 6): Figure 4). The same capabilities can be used to dynamically control and groom individual base station traffic on a call-by Broadcast layer is the core satellite-based broadcasting syscall basis. This technology is particularly suitable for situations tem. It gathers content from the various content providers in which many remote base stations (e.g. 100) each handle only and broadcasts it over the satellite and the terrestrial limited traffic (e.g. less than 2 Mbit/s), as is often the case in repeater network to the SDR receivers, which deliver the rural and low density areas. Other wireless access solutions, broadcast content to the applications. such as Unlicensed Mobile Access (UMA), are also candidates Point-to-point communication layer provides bidirecfor future broadband backhaul by satellite. Users equipped with tional communication between two remote points. It is a UMA / GMS hybrid phone would progressively benefit from basically a GSM, GPRS or future 3G network. cellularWiFi convergence over satellite. Application layer carries all the user services, relying on the other two layers for delivery. On the user terminal side, appli3G and beyond mobile satellite systems architectures cations can access content from both the SDR receiver and So far, mobile satellite systems have only been able to operthe cellular modem. ate with specific terminals using standards that were not compatible with terrestrial cellular systems. This applied both to system architectures based on low earth orbit satellites (e.g. 50 satellites or so) providing global coverage and to large geoFigure 6: SDR end-to-end communication architecture stationary satellite solutions (e.g. one or two Big GEO satellites) covering a continent, such as North America or Asia. SDR Broadcast Recent satellite improvements mean that Big GEO soluLayer tions can work with standard mobile handsets, making such solutions more user-centric (see Figure 5). Furthermore they are much less capital intensive to build and launch, and therefore less risky to deploy. Typical mobile satellite systems for 3G and beyond can now implement two user connection modes: Content Aggregation
Figure 5: Mobile satellite systems adoption of standard user terminals
Server
Direct connection from satellite to terminals. This is mainly for users who spend most of their time outdoors. Indirect connections, which require the satellite and a complementary terrestrial wireless network to access users in dense urban areas and/or deep indoors. The large antenna size of Big GEO systems (typically 25 m diameter) means that standard size user terminals with a 2 W transmission capability can transmit data on the uplink with an average bitrate sufficient to support interactive asymmetric broadband services. Satellite digital radio broadcasting services to vehicles In the US, two Satellite Digital Radio (SDR) systems have been operating since 2001. Today XM Radio serves more than
Terrestrial Repeater
SDR Receiver
IP Network
Application 1 Application 2 Application N
Application Layer
Content Providers
GSM/GPRS Modem
GSM/GPRS Network
Point-to-Point Communication Layer
www.alcatel.com/atr
125
The choice between the broadcast and point-to-point communication layers will depend on each service and on the operational conditions for each user. Any service requiring data originating from the user terminal will use the point-to-point communication layer. On the downlink path (service provider to users), the choice will be made according to the cost of delivering the content: information intended for many users will be broadcast, whereas information directed at only a few users will be provided through the communication layer. The choice will also be based on the time taken to deliver the information. The following is typical user information: Radio content: to be provided through the SDR broadcast layer. International or national traffic information, and large city traffic information will be provided through the SDR satellite broadcast layer. Local city traffic information will be provided through the terrestrial network in the city. Information sent to a closed user group will be provided through the SDR broadcasting layer. A carousel protocol will be used to ensure that content is successfully delivered. This protocol checks that content has been correctly received by the user. Depending on the importance of the content, the user terminal will automatically send an acknowledgment. At the end of a predetermined time, a user terminal that has not yet received the information may be updated directly using the point-to-point communication layer.
dards. Finally, the key requirement is that services should be independent of the technology (terrestrial and satellite), enabling users to communicate wherever they are and in any situations they might experience (see Figure 7). Support for broadband IP by satellite is certainly the right way to achieve this user-centric objective.
Figure 7: SDR end-to-end communication architecture

MSPs Telco / ISP/ MSOs Broadcasters
Terrestrial coverage
3G / MBMS
ADSL /Cable DOCSIS
DVB-T / DAB
Coverage extension via satellite
S-DMB
DVB-RCS +DVB-S2
DVB-S/ S-DRB
Dedicated terminals
Broadband User MSP: Mobile Service Provider MSO: Multi Service Operator (Cable Operators)
Extending Reach by Satellite to BroadbandDisconnected Users: Converging Standards?

Historically three distinct communication modes have coexisted independently to deliver services to users; distinct non-converging standards have been defined to support these media and services. Today, mobile service providers are focusing on the 3G standard and its definition of the Multimedia Broadcast / Multicast Service (MBMS). The Satellite Digital Multimedia Broadcast solution (S-DMB) currently being defined by Alcatel Space will extend 3G / MBMS services to users who cannot receive them at present. Consequently, users moving from terrestrial 3G coverage to satellite 3G coverage will only need to use a common 3G terminal. Current evolution of the 3G standard will see the deployment of High Speed Downlink Packet Access (HSDPA) and, ultimately, Beyond 3G / Fourth Generation (B3G/4G) services. Regarding fixed networks, telcos, ISPs and cable operators (multiple service operators) are concurrently deploying Asymmetric Digital Subscriber Line (ADSL) and cable / DOCSIS (Data Over Cable Service Interface Specification) to transport both IP and TV services. To extend the reach of these services, the satellite DVB-RCS standard has been defined to offer the same level of service as terrestrial systems, with an identical user experience for people not connected via DSL. Similarly, there will soon be a need to consider providing support for WiMAX (IEEE 802.16) via satellite. Terrestrial TV broadcasters and audio broadcasters have based their digital solutions on the DVB-T (Terrestrial) and the Digital Audio Broadcast (DAB) standards, respectively. The satellite counterparts are the DVB-S (Satellite) and S-DRB (Satellite Digital Radio Broadcast) stan126
Combined IMS / LBS Satellite Enhancement Architecture

A promising way to deliver new integrated solutions via satellite is the IP Multimedia Subsystem (IMS). This architectural framework, which has been designed to support real-time multimedia services, such as enhanced messaging, multimedia conferencing and video-based services, can be adapted to hybrid satellite / terrestrial solutions. IMS also enables new services to be introduced rapidly through the use of a standard service creation and service delivery environment. Moreover, the combination of IMS with satellite Location-Based Services (LBS) will make it easier for users to recover their full service preferences from anywhere in the network (home network coverage area or extended satellite coverage area). Accurate user presence information will mean that the IMS server can be in a better position to select the most appropriate communication infrastructure when establishing the requested user session. In this situation, attractive new services will become available to mobile users. For example, in addition to IMS user context data, locationaware applications will deliver associated map and direction data directly to the user handset. Telecom operators operating the emerging Assisted Global Positioning System (A-GPS) location servers (see Figure 8) should be in a privileged position to provide users with the appropriate communication environment whether they are on the move, in the office or at home. This solution should benefit from the existing European Geostationary Navigation Overlay Service (EGNOS) satellite augmentation system which provides reliable, high accuracy user positioning well before the full GALILEO satellite system becomes operational by 2009.
www.alcatel.com/atr
With this objective in mind, the European LIAISON project, which is led by Alcatel Space, has the aim of developing and providing turnkey location-based services for a wide range of mobile user communities.
Figure 8: Accurate user positioning based on A-GPS principles
GPS constellation and EGNOS-WAASMSAS satellites
> > >
http://liaison.newapplication.it: LIAISON European project. http://www.esa.int/export/ esaSA/navigation.html http://europa.ev/int/comm/dgs/ energy-transport/galileo/index_ en.htm
1- Satellite information is collected by the network periodically 1 Satellite Reference Receiver IMS User Data Server Enhanced A-GPS Server 3 3 A-GPS Capable Handset 3
2- On request, enhanced GPS information is transmitted to the user 2 4 2 4
Conclusion
The combination of broadband Wireless satellite solutions with wireline and Network wireless solutions (DSL, WiFi, WiMAX, GSM) can provide users with no terrestrial broadband connection with the same user experience as if they were connected to urban broadband sites. With this objective, the development of turnkey hybrid solutions, including satellite backhaul and DSL, WiMAX, GSM or WiFi, to the next generation network infrastructure will extend the broadband reach of telecom and entertainment services to disconnected users. By complementing the IMS functionality, satellite locationbased services will also help select the best network media and the content service context that best fits the users preferences, for example, by determining the most efficient network (satellite or terrestrial) according to whether the user is at home, in the office, on the move, etc.
4- User position is transmitted to the network
3- User position is determined by the handset or by the network
Advantages of the satellite enhancement architecture include: Brings together multiple environments through its ability to simultaneously work with wireline, wireless and mobile networks. Incorporates intelligence within the edge network to manage, provision and bill a large number of users with a wide choice of individual services. Uses a new generation of service-aware management tools integrated within the service providers platform to sustain an improved service delivery solution. New satellite service-aware architectures satisfy the requirements for rapid migration to a user-centric environment, enabling service providers to deliver the full service flexibility expected by users. Satellite solutions can therefore contribute to the transition to a user-centric broadband world in which users can enjoy services and interact seamlessly over multiple networks using a variety of terminal types.
EGNOS
The European Geostationary Navigation Overlay Satellite Service (EGNOS) augments the GPS and GLONASS satellite navigation systems so that they can be used for safety-critical applications such as aircraft navigation. GPS data is collected by 34 Ranging and Integrity Monitoring Stations (RIMS) in 22 countries and sent every second to Master Control Centers (MCC) in the UK, Germany, Spain and Italy for the addition of accuracy data. This augmented GPS data is then transmitted via satellites, with support from several Navigation Land Earth Stations (NLES). EGNOS enables users to determine their positions to within five meters, instead of about 20 meters at present.
Galileo
The future European Satellite constellation for navigation (equivalent to the US GPS).
Alain Bertout is Mobile and LBS Marketing Director, Business Development and Strategy Marketing, at Alcatel Space, Colombes, France. (alain.bertout@space.alcatel.fr)
Jacques Couet is Director in the Business Development and Strategy Marketing Department of Alcatel Space, Colombes, France. (jacques.couet@space.alcatel.fr)
www.alcatel.com/atr
127
TECHNICAL PAPER
ANY SERVICE
S. Ooghe, N. Drevon, R. Siebelink
SUPPORTING QUALITY OF SERVICE IN BROADBAND ACCESS NETWORKS

An optimal user experience for triple-play services can be assured thanks to QoS enablers in fixed and wireless access networks.
ompared with existing high-speed Internet services, new broadband services, such as interactive television and multimedia telephony, must meet very different user expectations, such as high service availability, flickerfree video quality and smooth voice communication. To ensure that such a user-centric broadband experience becomes a reality, the broadband fixed and wireless access networks must meet a number of Quality of Service (QoS) requirements, including guaranteed throughput, and low delay, jitter and packet loss, for these new services. For fixed access networks, such as Digital Subscriber Line (DSL) and Passive Optical Networks (PON), the desired user experience can be achieved by combining QoS techniques in the access node, for example, the DSL Access Multiplexer (DSLAM) or Optical Line Termination (OLT), and the aggregation network. Traditionally this could be achieved using Asynchronous Transfer Mode (ATM) QoS techniques. However, in view of the high demand for bandwidth-intensive services, operators are increasingly migrating to Ethernet-based transport. Consequently, access nodes and aggregation networks are becoming packet aware, allowing a number of new QoS techniques to be supported. The QoS mechanisms used in the access and aggregation networks need to be tied into appropriate network planning and configuration, resource control and policy enforcement measures. Numerous solutions exist, but no single technique has emerged as dominant. This article highlights the role of network planning on the DSL line combined with QoS control in the aggregation network, as well as the role of the access node.
128
In wireless broadband access networks, the QoS techniques depend on the radio transmission technology, for example, Global System for Mobile communication (GSM) / General Packet Radio Service (GPRS) or Wireless Local Area Network (WLAN), as well as the switching paradigm (circuit or packet switched) and overall integration with the backhaul network. The Unlicensed Mobile Access (UMA) and Third Generation Partnership Project (3GPP) WLAN interworking architectures provide a combined fixedmobile QoS solution that can leverage the QoS mechanisms used in fixed networks.
Providing QoS in Fixed Access Networks

Broadband access network architecture Figure 1 illustrates a typical broadband access network architecture ready
for triple-play services. It shows the DSL- or PON-based access network, the Ethernet / VPLS (Virtual Private LAN Service) based aggregation network and the Internet Protocol / Multi Protocol Label Switching (IP/MPLS) based regional network. The access nodes located in the central office are typically DSLAMs or OLTs, while the outdoor equipment generally consists of Very high speed DSL (VDSL) remote units. A Broadband Remote Access Server (BRAS) in the regional point of presence provides connectivity to Internet Service Providers (ISP). Additionally, one (or more) service routers can be used to provide video or Voice over IP (VoIP). To assure a good user experience, these services must be delivered with the appropriate level of QoS throughout the network, while optimizing the bandwidth efficiency on the first mile to
Figure 1: Access network architecture
Customer Premises
Central Office
Regional PoP
Ring/mesh/star Home Network ISP BRAS
Home Network
Remote Access Node DSL Modem Access Network Access Node Aggregation Network Service Router
WWW
Video Server Regional Network
PoP: Point of Presence
www.alcatel.com/atr
reduce operating costs. Each network element must differentiate the various application flows. This can be achieved by classifying the flows at the edges of the network (i.e. in the first access or aggregation node and in the application servers) into a limited set of aggregate flows characterized by QoS marking, such as the IP DiffServ Code Point (DSCP), the Ethernet priority class (802.1p bits) or the MPLS Experimental bits (E-LSP). These network elements provide a per subscriber QoS; the network elements in the aggregation network and regional network provide queuing and scheduling for the aggregate flows. QoS in the access node As it is the first network element under the full control of the access provider, the access node has a key role in providing QoS for upstream traffic. Additionally, the access node has a full knowledge of the access line characteristics for downstream flows and can therefore prevent downstream congestion of high priority traffic. A variety of techniques are used to optimize network load, application performance and fairness among applications. Upstream QoS Traffic classification and filtering is achieved using Access Control Lists (ACL), which help identify the traffic streams that require QoS treatment. It is expected that these ACLs will become increasingly important and will include not only ATM and Ethernet-based classification, but also IP and port-based classification. Filtering also enforces security policies by removing fraudulent traffic that could potentially compromise the overall network behavior. DSCP or p-bit (re)marking to values that can be trusted by the network elements in the aggregation network when providing per-flow QoS processing. The DSCP is typically used by QoS-enabled IP service routers, whereas the p-bits can be used by QoS-enabled Ethernet/ VPLS switches. If the DSL modem or home gateway has performed a traffic classification process, then the access node can use the DSCP or p-bit marking to perform per-flow QoS processing as well as remarking, if required. Otherwise the access node traffic filtering process classifies the traffic into QoS subflows, and then (re)marks the packets accordingly.
Ingress policing enforces the traffic contracts that specify how much traffic users can send to the network. A policer may apply to an entire access line or to a set of QoS subflows matching a multi-field classifier. Policing may take into account the drop precedence coded in the DSCP or p-bits; out-of-profile traffic can be subject to packet drop, or can be remarked with a higher drop precedence. Traffic forwarding to the right egress interface can be based on several fields; typically the destination Media Access Control (MAC) address and/or the destination IP address are used. Per-QoS class queuing and scheduling on the access node uplink(s) based on the DSCP or p-bits: a set of egress queues is present and a QoS class is mapped to a queue. The scheduling mechanisms determine the exact treatment given to packets in different queues. Different schedulers (e.g. strict priority or round robin) can provide different functionality on the same set of queues. Downstream QoS In the downstream direction, frames arrive at the access node with DSCP or p-bits that are properly marked by the application servers or service routers. The latter also perform ingress policing, which implies that the access node does not need to do this. After the forwarding decision, the following steps can be performed: Egress rate limiting: Similar to ingress policing, this is used to enforce traffic contracts that are associated with the services to which the users have subscribed. Per-QoS class queuing and scheduling on the access lines based on the DSCP or p-bits: downstream it is crucial to use multiple QoS queues per access line and to map traffic to a specific queue based on the DSCP or pbits. The scheduling mechanisms combined with the use of queue management features, such as (weighted) random early discard, determine the exact treatment received by packets in the various queues. QoS in the aggregation network The aggregation switches and serviceaware edges provide capabilities similar to those described above. It is important to note that once per-subscriber QoS has been enforced, the network elements deeper in the network can perform traffic scheduling according to the QoS class
using different DSCP or p-bits, different VLANs, VPLS instances or a combination of these. Specifically, the use of VPLS in the aggregation network provides key benefits from the perspective of traffic engineering the available aggregate resources for different service classes [1,2].
Network Planning
The QoS features in the access and aggregation networks ensure that, at times of congestion, best-effort traffic is dropped first so that higher priority traffic flows (e.g. voice or video) Access Control are not affected. List (ACL) A Besides these QoS sequence of features, good patterns to match network planning packets from a is needed to traffic stream, and ensure that no the associated high priority packfiltering actions to ets are dropped. be taken when a Network packet matches a planning certain pattern. includes partiPatterns could be tioning the availbased on fields able bandwidth such as the on the various Ethernet header, IP links into a header or bandwidth UDP/TCP port budget for each numbers. QoS class. If bandwidth guarResource antees are to be Admission given for differControl (RAC) ent traffic classes, Actions taken by then the bandthe network to width budgets of determine whether these classes a new flow can be must not exceed granted the the planned requested Quality bandwidth of Service (QoS) budgets. This is without impacting enforced using earlier guarantees, aggregate policand to block it if ing per QoS class, not. and resource admission conRound Robin (RR) (Weighted) trol techniques. A technique for Figure 2 scheduling multiple shows an examtraffic streams on a ple partitioning common network for voice, video link. Typically, RR is and data traffic used to provide a with a minimum portion of the guaranteed bandwidth of a bitrate. Bestnetwork link to effort and coneach traffic stream, trolled-load trafthereby providing a fic could be means to achieve assigned to the differentiated QoS. same partition, provided the
129
www.alcatel.com/atr
scheduling process is capable of assigning a fair share of the partition to each service class. Specifying a non-zero minimum guaranteed bandwidth budget for controlled-load and/or best-effort traffic ensures that these classes are not be starved out by voice or video traffic. When performing network planning on DSL lines, the safest approach is to base the bandwidth budgets on the minimum guaranteed DSL rate. If the rate turns out to be higher, then this excess bandwidth can be used by best-effort applications, such as web traffic. Similar network planning is undertaken in the aggregation network. The available bandwidth is partitioned per QoS class, and possibly even per service provider (or business customer). Each service provider will be subject to a traffic contract with the access provider, specifying how much voice and video traffic it can generate. This contract, which is part of the Service Level Agreement (SLA) with the access provider, generally reflects the various bottlenecks in the network.
Figure 2: Network planning on the DSL line
Budget for Voice Traffic (e.g. 90 kbit/s for one G.771 voice flow)
DSL Bitrate (e.g. 3.7 Mbit/s)
Budget for Video Traffic (e.g. 3 Mbit/s for one video flow)
Budget for Controlled Load and Best Effort (e.g. 128 kbit/s minimum guaranteed controlled-load and best-effort traffic)
Minimum DSL Bitrate (e.g. 3.5 Mbit/s)
Resource Admission Control

Following network planning, mechanisms are needed to limit the total number of video streams on the access line and in the aggregation network. Resource admission control checks whether a new service can be supported by the network. A typical approach for triple-play services is shown in Figure 3. Resource admission control is handled by a QoS control component, which performs a number of checks for each link in the access and aggregation networks. There
are various ways to implement this function, ranging from a distributed approach similar to the integrated services model using in-band QoS signaling, to centralized QoS control for access and aggregation [3]. A centralized approach assumes a full knowledge of the services provided to the customer and network planning based on a knowledge of the services, including the bandwidth budgets on the access lines. It makes sense to integrate a QoS control component in the access node to perform resource admission control for an access line, and to perform QoS control for the aggregation network elsewhere. In the case of services that rely on multicasting (e.g. broadcast TV), the access node is the first network
Figure 3: Resource admission control for unicast and multicast services
Customer Premises
Central Office
Regional PoP
SIP
Application Server
IGMP
element to receive multicast join and leave messages and perform traffic replication. Since it has full information about the access line bitrates, it can prevent subscribers from joining a multicast stream if there are insufficient resources on the access line. This ensures that the qualities of other multicast streams on the same access line are not affected. Subscribers are likely to request subscriptions to multiple services. This poses a challenge when it comes to coordinating the admission control process in the access node with QoS control deeper in the network, as these are not normally aware of each other. The access node is likely to become more involved in this process in the future. Next to performing admission control, there is a need to enforce the decision made by the QoS control functions. Enforcement needs to take place at the edges of the access and aggregation networks at points where the user experience would otherwise be affected. In the case of downstream unicast services, enforcement could be performed at an application gateway (e.g. IP edge router). In the case of multicast services, the access node controls which multicast streams are replicated on access lines. In the upstream direction, policy enforcement can also be performed in the access node or deeper in the network, depending on where the bandwidth bottleneck occurs.
Multicast Service Control
Unicast Service Control
Providing QoS in Wireless Access Networks

In wireless broadband access networks, the way in which QoS is provided depends not only on the service, but also on the radio transmission technol-
IGMP: Internet Group Management Protocol
SIP: Session Initiation Protocol
130
www.alcatel.com/atr
Figure 4: UMA architecture

AAA Server
Generic Access Network Up interface
GSM/GPRS Core Network
A Public Internet SGW GANC GSM Radio Access Network MS BTS Carrier Private Network BSC MGW
MSC/VLR
HLR
AP
Gb A Gb
GSM/GPRS Transport Network
SGSN
tion-based services, are provided in a seamless manner irrespective of the radio technology. When a user with a dual-mode handset moves within the range of an unlicensed network (WiFi), the mobile contacts the GANC over the IP access network. Signaling and user data are routed through the GANC using a single IP Security (IPSec) tunnel that provides data integrity and ciphering. 3GPP specifies four classes of traffic: conversational, streaming, interactive and background: Conversational traffic requires a guaranteed bitrate with low latency. Streaming traffic needs a guaranteed bitrate with loose latency. Interactive and background traffic classes (in the packet domain only) correspond to bursty traffic with various priorities, but without a guaranteed bitrate. After the QoS required over the GSM/GPRS Radio Access Network (RAN) has been determined according to the requested service and the operator policy, the GSM/GPRS core network requests the establishment of a suitable bearer to the mobile station. Specifically for UMA, the GANC determines the appropriate QoS in the RAN based on the traffic class and then assigns the appropriate DSCP marking. The end-to-end QoS model for UMA is shown in Figure 5. UMA specifies how the DSCP of an IP packet is processed by the Security Gateway (SGW) and at the mobile station to ensure QoS both along the IPSec tunnel and beyond. Downstream, the GSM media gateway for the circuit-switched domain and the SGSN for the packet-switched
AAA: Authentication, Authorization & Accounting AP: Access Point BTS: Base Transceiver Station HLR: Home Location Register MGW: Media GateWay
MS: Mobile Station MSC: Mobile Switching Center SGSN: Serving GPRS Support Node SGW: Security GateWay VLR: Visitor Location Register
ogy GSM / GPRS, Universal Mobile Telecommunication System (UMTS), etc and the wireless access architecture. To date, two architectures have been specified that allow GSM/GPRS services to be carried over packet-based radio technologies and a fixed backhaul network, such as DSL. Both architectures reuse the QoS principles network planning and admission control described previously. Unlicensed Mobile Access UMA is a flexible fixed/mobile convergence solution that is intended primarily to extend GSM/GPRS coverage in the residential (home and public hot spot) and Small Office / Home Office (SOHO) markets. The UMA specifica-
tion, which has been transferred to the 3GPP, uses WiFi or Bluetooth access combined with fixed backhaul (e.g. based on DSL). This provides GSM/GPRS service support in the WLAN as well as seamless hand-off and roaming within the cellular network. As shown in Figure 4, the solution is based around a Generic Access Network Controller (GANC), which mimics the role of a GSM Base Station Controller (BSC), connected to a standard GSM/GPRS core network using standard interfaces. Voice and data GSM/GPRS services, including the Short Message Service (SMS), Multimedia Messaging Service (MMS), Wireless Application Protocol (WAP) and loca-
Figure 5: QoS handling in UMA
Outer IP header QoS (DSCP/ToS) 802.11e/BW 802.1p/BW
Inner IP header QoS (DSCP/ToS) 802.1p/MPLS/etc.
Home Gateway
Aggregation Network
Internet Core Data Network
MS
AP
RG
BRAS
SGW
GSM MG (CS)/ GPRS SGSN (PS)
BRAS: Broadband Remote Access Server CS: Circuit Switched
MPLS: Multi Protocol Label Switching PS: Packet Switched
RG: Routing Gateway ToS: Type of Service
www.alcatel.com/atr
131
domain mark downstream IP packets with DSCP. The SGW encapsulates the original IP packet into the IPSec tunnel towards the mobile station and builds the IPSec DSCP from the original DSCP. The original DSCP is retrieved from the encapsulated IP packet at the mobile station side. Upstream, the mobile station copies the downstream outer IP header DSCP to the upstream outer IP header DSCP for QoS between the mobile station and the SGW. It also copies the downstream inner IP header DSCP to the upstream inner IP header DSCP for QoS beyond the SGW. To guarantee QoS across Ethernet access and aggregation networks, the routing gateway maps the IP DSCP to the appropriate Ethernet QoS class. Additionally, the WLAN access point maps the 802.1p bits to the appropriate 802.11e WiFi multimedia QoS class in order to provide differentiated channel access, resulting in controllable and predictable transmission times and throughputs. 3GPP WLAN interworking 3GPP has specified how 3GPP systems and WLANs can interwork. The aim of the interworked WLAN is to extend 3GPP packet-based services to the WLAN access environment. Unlike
the UMA architecture, which is mainly designed for circuit-switched services, the interworked WLAN architecture is designed for packet-based services by connecting the WLAN to the same external IP network used by the GPRS access network. As shown in Figure 6, a WLAN is connected to an external IP network via a WLAN Access Gateway (WAG) and a Packet Data Gateway (PDG). The WAG, which is located in the visited network, acts as a firewall, enforcing routing of packets through the PDG and providing legal interception and charging. After authentication by the AAA server, an IPSec tunnel is established between the mobile station and the PDG in a similar way to UMA. 3GPP is on the way to standardizing QoS support over the interworked WLAN. This work might affect the IEEE 802.11e standard, amending the 802.11 WLAN standard with MAC enhancements for QoS. Since interworked WLAN services and UMA services are both 3GPP services, and therefore based on the same QoS principles, it is expected that they will use similar mechanisms. User signaling will request and negotiate the service, while network signaling is used to request the appropriate bearer to the access network. This is achieved
using the GANC for UMA and the PDG/WAG for the interworked WLAN. Once this has been done, the QoS techniques outlined in the UMA architecture and the DSL access network can be used unchanged.
Conclusion
To ensure a good user experience for real-time and streaming services, broadband fixed and wireless access networks must comply with a number of QoS requirements. The networks must be able to classify subscriber traffic per application type and provide differentiated traffic treatment before aggregating it into a limited set of service classes. Whereas per-subscriber traffic treatment will be performed at the network edges, aggregation and regional networks can provide queuing and scheduling for the aggregate flows. In DSL access networks, the DSLAM is aware of all the access line characteristics and can thus play an important part in preventing downstream congestion of high priority traffic and in enforcing per-subscriber traffic policies. Moreover, appropriate network planning on the DSL line and the aggregation network is required to ensure that no high priority packets are dropped or delayed during periods of
Figure 6: Interworking WLAN 3GPP packet-switched services provided from the 3GPP home network
Visited Public Land Mobile Network
3GPP Home Network Public Internet Wa 3GPP AAA Proxy Wm Wp WAG PDG Packet Switched Domain 3GPP AAA Server Wx HLR/HSS
WLAN Access Network Ww AP
Wn
Wi Packet Data Network
MS Gb
Gn GGSN SGSN BTS BSC A Nc/Nb MSC/VLR
Gi
GSM Radio Access Network
GMSC Circuit Switched Domain
GMSC: Gateway Mobile service Switching Center
HSS: Home Subscriber Server
132
www.alcatel.com/atr
congestion. This goes hand in hand with the use of resource admission control to check whether a new service can be accepted by the network. Thus the DSLAM has an important role, particularly in the case of video services, such as broadcast TV. QoS techniques are equally important in wireless broadband access networks. The UMA architecture primarily allows fixed/mobile convergence for circuit-
switched services by combining QoS features in the wireless access network with QoS-enabled backhauling using, for example, DSL. Similarly, the 3GPPWLAN architecture provides a way to extend packet-switched 3GPP services over a WLAN. It is expected that the interworked WLAN architecture will use the same QoS techniques as those used in DSL access networks and the UMA architecture.
Sven Ooghe is responsible for access functional strategy and coordinating broadband architecture standardization work in the Chief Technology Office team, Access Network Division, in the Alcatel Fixed Communications Group, Antwerp, Belgium. (Sven.ooghe@alcatel.be)
References
[1] M. Sif, L. Newell: Optimizing the Broadband Aggregation Network for Triple-Play Services , Alcatel Telecommunications Review, 4th Quarter 2004, pp 406-412. [2] R. Hoebeke, R. Hoekman, S. Van den Bosh, M. Leo, A. Berthillier: Migrating Metro Ethernet to MPLS, Alcatel Telecommunications Review, 4th Quarter 2004, pp 424-432. [3] S. Ooghe, T. Pollet: Unleashing Advanced IP Services in DSL Access Networks, Alcatel Telecommunications Review, 2nd Quarter 2003, pp 160-164
Nicolas Drevon is an endto-end mobile network solutions architect in the Chief Technology Office of the Mobile Communications Group, Vlizy, France. (Nicolas.drevon@alcatel.fr)
Ruud Siebelink is Product Line Manager for the Alcatel 7302 ISAM in the Access Network Division, Antwerp, Belgium. (Ruud.siebelink@alcatel.be)
www.alcatel.com/atr
133
TECHNICAL PAPER
ANY SERVICE
R. Sharpe, D. Zriny, D. De Vleeschauwer
ACCESS NETWORK ENHANCEMENTS FOR THE DELIVERY OF VIDEO SERVICES

Enhancements are being researched that will improve both the quality of video delivered over networks experiencing congestion or transmission errors, and the video channel changing performance.
igital Subscriber Line (DSL) networks have been successfully deployed primarily to provide a high-speed residential Internet service. Now a number of factors are leading network access providers to expand their service offerings to include the delivery of next generation voice and an assortment of video services. The common network architecture, as described in DSL Forum TR-59 [1], utilizes the Point-to-Point Protocol over Ethernet (PPPoE) between a routing gateway in the home and the network providers or Internet service providers Broadband Remote Access Server (BRAS). The access network is unaware of what type of traffic is being sent through the PPPoE tunnel, and therefore treats it all as best-effort traffic. It is up to the BRAS and routing gateway to correctly schedule traffic to avoid congestion and any consequent uncontrolled packet loss. New IP-aware network architectures, such as those described in [2,3,4,5], provide traffic class and IP awareness using the Ethernet 802.1p bits, Ethertype, DiffGroup of Serv Code Points Pictures (GOP) (DSCP), or other A collection of indications to (consecutive) differentiate betimages which are ween traffic encoded as one classes. It is then entity. It starts with possible to treat a key image some classes of (referred to as an Itraffic differently frame), which is from other classes. encoded without In access netreference to any works delivering other images, and voice, video and ends just before the data services, next key frame. video streams generally repre-
sent the bulk of the traffic. All video packets are treated equally, regardless of their importance to the decoded video. However, some video packets may be much more important than others. If a video-aware DSL Access Multiplexer (DSLAM) could inspect these packets, determine their relative importance, and treat the traffic accordingly, then the overall delivery of video and other traffic could be improved. It is expected that DSL will benefit from ever-increasing video compression factors. However, this has the unwanted side effect of lengthening the time taken to change channels. This article presents
Figure 1: GOP structure
Forward Prediction
I B B P B B P B B P B B
Bidirectional Prediction a method for reducing the channel changing time, thereby enabling a higher degree of video compression.1
Video Compression Fundamentals

Video is composed of a sequence of pictures or frames. In the case of enter-
tainment quality video, the popular MPEG-2 and emerging H.264 (also known as MPEG-4 Advanced Video Coding or AVC) [6] techniques perform compression over a Group Of Pictures (GOP). A GOP is a sequence of three types of encoded frames: an Intra-coded (I) frame, Predictive (P) frames, and Bipredictive (B) frames. A representative GOP structure is shown in Figure 1. The GOP begins with an I-frame, which is encoded independently of any other frame. The I-frame is used to predict what other frames will be in the GOP. P-frames are encoded with reference to a past I- or P-frame and may be used as a reference to future P- or Bframes. B-frames are encoded with reference to a weighted average of two reference frames (I or P for MPEG-2; I, P or B for H.264). Both the encoder and
This work was partially funded by a research grant from the US Department of Commerce, National Institute of Standards and Technology (NIST), Advanced Technology Program [12]. Sarnoff Corp, Thomson and SBC are also members of the NIST joint venture.
134
www.alcatel.com/atr
Figure 2: Network Abstraction Layer (NAL) unit format
NAL Unit Header

forbidden_zero_bit
NAL Unit Payload
nal_ref_idc, signals relative transport priority of the NAL Unit nal_unit_type
(RTP/UDP/IP) [7]. NAL unit payloads of various types are encapsulated into NAL units; the NAL unit type and relative transport priority are indicated by the two NAL Reference ID (NRI) code bits in the single byte NAL unit header (see Figure 2). The relative transport priority can be used by a video-aware transport network to ensure that NAL units containing more critical payloads receive a higher quality of service than less-critical payloads.
DSL Performance
decoder predict what the frame currently being compressed will look like, based on past reconstructed frames (the encoder includes a decoder so that both the encoder and the decoder make an identical prediction). The encoder encodes the difference between the actual frame and the predicted frame. The prediction error is encoded and transmitted to the decoder and used with the decoders prediction to reconstruct the frame. This background shows that not all packets are equally critical to the quality of the reconstructed video. For instance, a defect in a reconstructed Iframe caused by a transmission error or packet loss will propagate to all other frames in the GOP and may be quite objectionable. A defect in a reconstructed P-frame affects fewer frames and is often noticeable, whereas a defect in a reconstructed B-frame generally affects only itself and may often be effectively concealed. H.264 is a significant improvement over MPEG-2 in several ways, including: Compression performance: H.264 is capable of 50% or greater bitrate savings for a similar degree of encoder optimization; this has been achieved by many cumulative improvements. Improved network friendliness to ease packetization, information priority control, and application to video streaming services. Enhanced error resilience and packet loss resilience tools. H.264 specifies a layered structure with a Video Coding Layer (VCL) efficiently representing the content of the video data, and a Network Abstraction Layer (NAL) to format the data and provide header information in a way that enables it to be conveyed over transport layers, such as MPEG-2 transport streams or Real-Time Transport Protocol / User Datagram Protocol / IP The data rate that can be supported over an access line decreases with increasing loop length. It is also a function of the wire gauge, the condition of the cables and splices, crosstalk and noise levels, and other factors. Figure 3 shows the simulated performance of Asymmetric Digital Subscriber Line 2+ (ADSL2+) and Very high speed Digital Subscriber Line 2 (VDSL2) over 26 AWG (0.4 mm) wire with a fairly significant level of self-crosstalk and noise. The performance on outside plant is often better, but can also be worse than this simulated performance.
Crosstalk An unwanted energy transfer between a disturbing pair in a telephone cable, or a disturbing channel of a transmission system, and disturbed pair or channel. It is called either far-end crosstalk (FEXT) or nearend crosstalk (NEXT) according to the location of the measurement point with respect to that of the disturbing source.
Figure 3: Simulated DSL performance
DSL Performance
25
20
15
10 VDSL2 Downstream ADSL2+ Downstream VDSL2 Upstream 0 0.00 1.00 2.00 3.00 4.00 5.00 6.00 ADSL2+ Upstream
Using H.264 video compression, three channels of standard definition TV, one channel of high definition TV, a few channels of Voice over IP (VoIP), with some capacity left over for high-speed Internet access requires approximately 18 to 24 Mbit/s to a user. As can be seen from Figure 3, 20 Mbit/s can be provided over about 3000 ft (1 km) using VDSL2 or about 4500 ft (1.5 km) using two-pair bonded ADSL2+ (10 Mbit/s per pair). Clearly this distance is too short to provide the required suite of video services to most users served directly from the central office.
Data Rate (Mbit/s)
Multi-service Access Networks

DSLAMs have traditionally been deployed in central offices, since the ability to provide high-speed Internet service to a large number of users from a central location is attractive from the
135
Loop Length (thousands of feet)
www.alcatel.com/atr
viewpoint of equipment and operating costs. However, the distance from central office to user is often too long to support the desired suite of video and other services. Consequently, to achieve the high data rates required, it is necessary to deploy DSLAMs deep into the network to considerably shorten the distance to the user. There are a limited number of points between the central office and the user where a carrier can conveniently gain access to a twisted wire pair in order to couple DSL to it. These points are a consequence of the outside plant engineering rules established long before DSL networks were envisioned. In the United States, carriers define Carrier Serving Areas (CSA) based on loop
Figure 4: Multi-service access network
Internet
Central Video VoD Service Server Mgmt
BRAS Distributed VoD Server
DA Carrier Serving Area Broadcast TV Head-end
Packet Core Network

VoD: Video on Demand
Packet Aggregation Network
DA Distribution Area
Alcatel 7330 remote unit
ther improve the delivery of triple-play services.
Video-aware Traffic Management

Figure 5 2 compares the consequences of fairly discarding video packets with selectively discarding video packets for H.264 compressed video sequences. The images are of frames captured from decoded video after an equal number of bits have been discarded. The image on the left is of a captured image following the discarding of B-frame packets, while the captured image on the right is after discarding P-frame packets. The conse2
resistance, which is related to wire gauge and loop length. These CSAs are divided into Distribution Areas (DA), primarily based on cable route and management considerations. At the CSA and DA boundaries, known as the Feeder Distribution Interface (FDI) or Service Area Interface (SAI), a crossconnect cabinet is generally installed for cable rearrangement and tapering (i.e. to accommodate the increasing number of twisted pairs per home closer to users homes). DSL can be coupled to the twisted wire pairs at these locations. Figure 4 shows a multi-service access network with DSLAMs deployed at the FDI locations. Assuming a well engineered core and aggregation network, the volume of traffic that can ultimately be delivered to a user is constrained by the DSL line rate. Deploying DSLAMs as deep as is practical significantly addresses this constraint. Improving the traffic management at this point of congestion can fur136
Courtesy Thomson
quence of intelligent packet discarding is negligible while the consequence of random packet dropping is quite objectionable and lingers until the beginning of the next GOP. Generally, a video stream is encoded with a constant bit rate, or with a variable bit rate constrained by peak and average rates. Encoders can also be implemented with additional rate constraints and still comply with compression standards. For example, an encoder could limit the peak rate of a base layer (the I- and P-frames) and allow more variability of an enhancement layer (the B-frames). Connection admission control or policy driven methods can ensure that there is sufficient capacity on each DSL line to carry the video base layer of the selected video streams in addition to voice and other high priority traffic. The video enhancement layer of the selected video streams and lower priority traffic could be scheduled with a
Figure 5: Intelligent (left) versus random (right) packet dropping (Source: Thomson)
www.alcatel.com/atr
Figure 6: Fast channel change approach
Encoder
Channel 1 Channel n
Core Network
Aggregation Network
DSLAM
RG
Set-Top Box
Viewer
Viewing FR Ch1
Leave FR Ch1 Join RR Ch n Join FR Ch n
Change to Ch n
Decode RR Ch n Leave RR Ch n Decode FR Ch n
Viewing RR Ch n Viewing FR Ch n
Viewing FR Ch n
Full Resolution (FR) Stream Reduced Resolution (RR) Stream
lower priority, allowing a degree of overbooking without compromising the most important traffic. In IP DSLAMs, such as the Alcatel 7302 and 7330, traffic can be treated differently based on the 802.1.p bits. The .p bits may be trusted or be determined by inspecting fields such as the Ethertype or DSCP field in the arriving packets. In a video-aware DSLAM, this method can be extended to include inspecting the NRI bits in the NAL unit header of the arriving video packets. It is prudent to use the NRI bits rather than trusting the .p bits as .p bits set by an encoder may be altered as packets traverse the network.
Fast Channel Changing

A significant factor in the users experience is the channel changing time. A viewer expects a new channel to appear very rapidly after the channel has been selected. The acceptable range is about 500 ms or less. It has been observed that extending the GOP period from 0.5 s to 2 s improves the coding efficiency by about 15 to 20%. However, as decoding can only start after an I-frame has been received by a decoder, improving the coding efficiency by extending the GOP period would unacceptably degrade the channel changing performance. A solution currently being investigated [8] is to associate a low bitrate (reduced temporal and spatial resolution) stream with
each channel. This stream will be encoded with I-frames that occur frequently (2 to 4 times a second) so that the stream can be quickly decoded and displayed. The capacity required for the reduced resolution stream would be less than the savings achieved from the longer GOP period. The fast channel change approach is illustrated in Figure 6. When a set-top box requests a channel change, the DSLAM delivers both the full resolution stream and the associated reduced resolution stream, which could be quickly decoded and displayed. Soon after the full resolution Iframe is received, the set-top box will display the full resolution stream and signal the DSLAM to stop sending the reduced resolution stream.
Visually the new channel will initially appear as a soft (i.e. low resolution) image, which will quickly sharpen. While the reduced resolution video may be detectable immediately following a channel change, having the new channel appear quickly is much more important to the viewers experience. The reduced resolution streams could also be used for Picture-In-Picture (PIP) and for navigational (channel selection/changing) purposes.
Forward Error Correction

A DSL line is prone to noise from stationary noise sources, such as crosstalk from signals on other twisted wire pairs in the same cable, and non-stationary noise sources, such as telephony onhook/off-hook transients, and from many external noise sources. Noise causes transmission errors and consequently packet loss. To protect against transmission errors, DSL uses trellis coding and Reed-Solomon Forward Error Correction (FEC). The latter can be achieved with or without interleaving, a technique that disperses bursts of errors so that they can be corrected more effectively. The non-interleaved (or fast) mode is effective against stationary noise sources; it is considered sufficient for services that can tolerate occasional packet loss, such as high-speed Internet access. However, more error sensitive services, such as video, require further error protection.
137
Stationary noise Noise (e.g. on the DSL line) for which the statistical properties (average noise energy, frequency spectrum) do not change over time. Although the statistical properties do not change, the noise signal itself is highly random. Trellis coding Trellis coding is convolutional coding combined with a bit-to-symbol mapping that achieves coding gain with no bandwidth expansion.
www.alcatel.com/atr
Figure 7: Mean time between visible distortions (top) and additional delay due to FEC (bottom) as a function of the overhead bitrate due to FEC.
Time between visible distortions [h]
96 Standard Definition TV (1.5 Mbit/s) 72 High Definition TV (8 Mbit/s)
48
24
0 0.0% 2.0% 4.0% 6.0% 8.0% 10.0%
Overhead bitrate due to FEC [%]
Forum Code of Practice 3 (CoP3) [10]. RFC2733 calculates an FEC packet for a set of fixed length data packets. Typically the FEC packet contains only the parity of the other packet payloads. Collecting the set of packets introduces latency and the FEC packets themselves introduce overhead. This scheme allows the loss of a single packet in the set to be recovered. Figure 7 illustrates the performance of the RFC2733 scheme for standard definition and high definition TV transported over MPEG-2 transport streams (according to RFC2250 [11], Section 2) over a 20 Mbit/s DSL line in interleaved mode with a bit error ratio of 10-7. From this figure, it can be seen that in the case of high definition TV, the average time between visible errors is greater than 12 hours, although this comes at the expense of 2.5% overhead and an additional delay of just over 50 ms.
Unequal Error Protection

Additional delay due to FEC [ms]
300 Standard Definition TV (1.5 Mbit/s) High Definition TV (8 Mbit/s)
250
200
150
100
50
0 0.0% 2.0% 4.0% 6.0% 8.0% 10.0%
Since the loss of some video packets results in considerably more visible defects than others, and when the importance of a packet can be easily determined, error protection can be applied unequally so that the most important video packets receive the most protection. For example, packets containing Bframes may not warrant additional error correction beyond that already provided by the DSL link layer, whereas packets containing I- and P-frames could be protected in accordance with RFC2733. This results in either less overhead for a given level of protection, or greater protection for a given overhead.
Conclusion
The delivery of video services will have a major impact on the access network architecture, equipment design and equipment placement because peak and sustained traffic levels are dramatically higher than required for highspeed Internet access. Sensitivity to packet loss is also greater. Non-blocking architectures, the placement of IP-aware DSLAMs closer to users, and DSL link layer FEC improvements all contribute to making an access network capable of supporting voice, video and data services. Improved video compression and reduced sensitivity to packet loss enable more channels of higher quality to be delivered over longer distances than would otherwise be possible. The H.264 video compression standard provides a mechanism for identifying the importance of a video packet. A video-aware DSLAM can differentially treat video packets in a manner similar to that in which Ethernet and IP net-
Overhead bitrate due to FEC [%]
Additional error protection can be provided in various ways, which can be used alone or in combination. First, the noise margin used by the DSL modem when training can be increased. This is effective against stationary noise, but is of limited use against impulse noise. Increasing the noise margin reduces the DSL line rate at a given distance. Second, the interleaved mode Impulse noise can be used to Noise consisting of combat bursts of random occurrences errors. Increasing of energy spikes the interleaving with random depth increases amplitudes and burst error prospectral content. tection, but at the cost of increasing
138
transport latency. Increased latency can decrease the throughput of services that are sensitive to round-trip delay, such as Transmission Control Protocol (TCP) transfers. A third technique is to use application layer error protection, which enables applications to provide additional error protection as and when necessary. Some application packets could be given a greater degree of protection than others. As an end-to-end error protection scheme, this has the additional benefit of correcting errors in other parts of the network, such as the core and home networks. A simple FEC scheme is proposed in RFC2733 [9]; an extension is utilized to carry video traffic in the Pro-MPEG
www.alcatel.com/atr
works are able to differentially treat packets based on identified markings. While some degree of FEC is appropriate for all traffic on a DSL line, it might not be the most efficient way to provide sufficient error correction for the most sen-
sitive traffic. Additional application layer error correction of just the most sensitive packets may be more efficient and needs to be investigated further. A satisfactory user experience is crucial to the success of a video service. One
element of this experience is the time taken to change channels. Techniques are being investigated that improve the channel changing performance while allowing greater video compression.
References
[1] DSL Evolution Architecture Requirements for the Support of QoS-Enabled IP Services, DSL Forum TR-059, September 2003. [2] S. Ooghe, E. Six: New Service Architectures for DSL Networks, Alcatel Technology White Paper, November 2003. [3] J. De Jaegher, P. Vandaele, D. Chantrain, D. Davy: Multi-Service Ethernet Broadband Access Solutions, Alcatel Technology White Paper, September 2004. [4] M. Sif, L. Newell: Optimizing the Broadband Aggregation Network for Triple-Play Services, Alcatel Telecommunications Review, 4th Quarter 2004, pp 406412. [5] J. Segel: Engineering in Multi-Service Broadband Networks, Alcatel Telecommunications Review, 4th Quarter 2004, pp 413417. [6] ITU-T Recommendation H.264, Advanced video coding for generic audiovisual services, May 2003. [7] Wenger et al: RFC3984 RTP Payload Format for H.264 Video, February 2005 [8] J. Boyce, A. Tourapis: Fast Efficient Channel Change, Proceedings of ICCE, 2005. [9] J. Rosenberg et al: An RTP Payload Format for Generic Forward Error Correction, Internet Engineering Task Force RFC2733, http://www.ietf.org/rfc/rfc2733.txt, December 1999. [10] Pro-MPEG Forum: Transmission of Professional MPEG-2 Transport Streams over IP Networks, Pro-MPEG Code of Practice #3 Release 2, http://62.73.167.57/publications/pdf/Vid-on-IPCoP3-r2.pdf, July 2004. [11] D. Hoffman et al: RTP payload Format for MPEG1/MPEG2 video, Internet Engineering Task Force RFC2250, http://www.ietf.org/rfc/rfc2250.txt, January 1998. [12] Video-Enhanced Residential ADSL Broadband Technology, NIST ATP Award Number 70NANB3H3053.
Randy Sharpe is Senior Principal Engineer in the Alcatel Network Strategy Group, Raleigh, North Carolina, USA. (Randy.sharpe@alcatel.com)
Don Zriny is Senior Development Engineer and a Member of the Alcatel Technical Academy in Alcatel Research and Innovation, Plano Texas, USA. (Donald.Zriny@alcatel.com)
Danny De Vleeschauwer is a Research Engineer and Member of the Alcatel Technical Academy in the Alcatel Network Strategy Group, Antwerp, Belgium. (Danny.de_vleeschauwer@alcatel.be)
www.alcatel.com/atr
139
CONSOLIDATION
J. van Bogaert, Y. TJoens, J-P. Lartigue
FIXED ACCESS VISION

Guaranteed delivery of triple-play services making optimal use of valuable resources is the key driver moving fixed access forward as the world converges towards offering all services over ubiquitous broadband connectivity.
One-stop shopping The user needs to contact only one entity to cover all his or her needs for connectivity and services. When subscribing to broadband access, the user can order not only Internet access, but also voice and video services. Customer churn When the user has no close links with the network operator, he or she may be tempted by promotional offers to change providers. Network operators perceive this as customer churn, bringing additional costs in the form of multiple activation/ deactivation cycles. Service bundles increase customer loyalty and reduce churn. Deep fiber strategy The strategy of pushing the fiber termination deeper into the network, i.e. closer to the user. The most important approaches are Fiber to the Home and Fiber to the Node.
he user-centric broadband world is positioning broadband access as the universal service delivery method. The prime service is no longer voice, but a combination of Voice over Internet Protocol (VoIP), data (remote working, Internet access) and video (video on demand, TV broadcasting, etc). These services are conveniently bundled to offer users a one-stop shopping experience, also known as the triple-play offer. Triple-play deployment is regarded by many service providers as a means to sustain and increase revenue, compensating for the loss in revenue from voice-only services, and as a means to fight customer churn. Integration between services at all levels of the network is essential. The quest to deliver service bundles over the broadband infrastructure is leading to four types of transformation involving access networks. First, the new generation of services, such as TV and Video on Demand (VoD), require a deep fiber strategy across the service provider base. Second, the increase in service awareness and finely tuned policy enforcement are reflected in the adoption of more intelligent access devices. Third, investment in broadband networks that support carrier-grade VoIP is an opportunity for anticipating the obsolescence of the Public Switched Telephone Network (PSTN). Finally, the need to guarantee service continuity while on the move is pushing operators to converge the operation of their infrastructures, whether the service is provided over fixed or wireless access technologies.
Broadband Technologies
Recent years have been witness to a significant increase in the maximum bitrates for both wireline and wireless access. This has frequently been matched by a corresponding decrease in range on the first mile. Alternative broadband deployment methods are now becoming available. Digital Subscriber Line (DSL) technology has improved to the point at which operators can confidently use their copper networks to deliver top quality triple-play services. Wireless technology is also getting prepared. The success of the IEEE 802.11 (WiFi) standard for supporting broadband homes has created a momentum around affordable radio technology. WiMAX (IEE 802.16) is destined to be the carrier version as it can use licensed spectrum, and ensure quality and secure service delivery. Fiber to the Home (FTTH) is, for many operators, the target for their outside loop deployment. Figure 1 gives an overview of the investments needed in two situations, namely greenfield with a 100% take up rate and brownfield with a 40% take up rate. In general, greenfield situations require a higher investment because of the need to deploy fiber and/or copper outside cable plant. This cost is partly avoided in Fiber to the Node (FTTN) and fully in central office based brownfield deployments, as they make use of the existing copper plant. From this it can be concluded that FTTH it is a feasible option in areas without an existing copper infrastructure, or when the outside plant needs to be replaced. Most of the cost is for the civil works associated with trenching, so the cost difference between fiber and copper is relatively low. In most cases, however, the copper plant will be in good condition, making FTTN used with Very high speed Digital Subscriber Line (VDSL) or
140
www.alcatel.com/atr
FIXED ACCESS VISION
Asymmetric Digital Subscriber Line Figure 1: Bandwidth and investment per deployment scenario 2plus (ADSL2plus) the most suitable technology. Investing in FTTN might be Per user Bandwidth justified in its own right, or can be con(Mbit/s) sidered as a cost-effective step towards 30 FTTH. The FTTN approach introduces 1500 fiber step by step, ensuring that the investment in triple-play services is in 20 1000 line with the revenue generated. This 40% take up rate two-stage deployment (FTTN with (brownfield) 10 500 FTTH in new areas, then FTTH every100% take up rate where) can be implemented at virtually (greenfield) the same lifecycle cost. As a result, Increased Fiber Penetration Copper Loop mixed deployment of DSL and 4 km 2 km 1 km 0 km Length FTTH will probably be the prevailCentral office ing strategy in the coming years. deployment When fiber is deployed to within 100 ADSL2plus to 300 m of the user, Ethernet cabling FTTN VDSL2 becomes an option, especially for multitenant units where CAT5 wiring is FTTP PON already installed. FTTN and FTTH are both well aligned with the objective of moving TV services away from pure flows, using the techniques developed to create Virtual Private broadcasting to provide a richer experience that is more in line Networks (VPN), such as Virtual Private LAN Service (VPLS) with an individual users preferences. Compared with Hybrid and Multi Protocol Label Switching (MPLS). Mechanisms are in Fiber Coax (HFC) networks, DSL and FTTH have the advantage place to ensure end-to-end delivery. The network is therefore of offering bidirectional intelligent, that is, the user needs to interwork with the nethigh-bandwidth connecwork provider to ensure that he or she gets what is being charged tivity to each user. Profor. The network operator will deploy a service delivery platform viding the same service that will keep track of all the sessions in the network, at least over the cable infrafor the premium services that are being guaranteed and structure requires the charged. In addition, operators are creating added-value servcable segments to be ices to offer the user a richer set of services than is available on split into such small the open Internet, thereby ensuring customer loyalty. fiber-fed fragments that For the operators, the first model is unsustainable from a ecoit is unable to compete nomic point of view. If there are no service revenues to be earned economically. It is and only a flat connection fee, there is not enough incentive to invest expected that telecom in a broadband network that gives all users enough bandwidth. providers will increasHowever, it is Alcatels belief that the second model will allow ingly take advantage of this capability to offer a richer TV expetodays operators to benefit from a sustained advantage, as it rience to broadband users than they are getting from cable. This enables them not only to provide the same service set in terms will drive the wide adoption of broadband services to non-PC of functionality and price (cost structure), but also to compete users as well, leading to a situation in which operators can conon the basis of reliability, quality and ease of use. This model sider making broadband ubiquitous by connecting the remainenables network operators to bundle services with the broading narrowband-only customers to the broadband infrastructure. band network offer (triple play). It will convince network providers that there is enough money to be made and that large scale investment in broadband is justified to support VoIP, video, Triple-Play Delivery IP TV and other services. This represents a shift in value from It is clear that broadband for triple-play services is voice minutes to triple-play bundling and service integration. essentially a problem of bandwidth. It could be argued Apart from being lucrative, it can also be tailored to individual that the only thing expected of a broadband access network is user and service needs and expectations, leading to a more scalto deliver an Internet Protocol (IP) connection to users with an able and robust delivery of content and communication services. adequate bandwidth at the lowest cost. However, the telecoms To realize this added value, broadband access networks must world is looking at two models. offer more than just high-speed Internet access, but distinguish The first model sees the public Internet as the only network at the service level between best-effort services and premium that is needed. Provided enough bandwidth is provisioned, it guaranteed services. Broadband access resources balancing should be able to support all the desired services. However, this reduced latency with increased error robustness must be allonetwork is basically dumb, meaning that it is totally service cated on a per-service basis. This enables QoS-based service difagnostic. All the intelligence is implemented at the boundaries, ferentiation, but more intelligent service hooks1 will be needed, terminals or servers. Service charging is totally transparent to the network. Examples are Vonage and AT&T CallVantageSM, such as secure resource protection and resource control linked to dynamic sessions. which offer VoIP over any broadband line. The second model foresees a per-service segmentation of the A service hook is a system capability that allows an operator to support IP network. New segments are created next to best-effort IP
1
a service at a later date.
www.alcatel.com/atr
141
FIXED ACCESS VISION
Premium services Services for which users pay a charge on top of the basic subscription fee. Headroom Additional capacity that is available in the network to accommodate situations that are beyond what is normally anticipated. Identity spoofing Without the necessary precautions in the network, malicious users could mimic the identity of others. This can result in unwanted phenomena like theft of service or Internet attacks that cannot be traced back to the originator.
The ultimate bottlenecks in any deployment scheme will be the first mile and the aggregation network. The closer the network comes to the users, the less chance there is of relying on bandwidth sharing techniques, such as statistical gain and the use of broadcast trees in which the same traffic flows are sent to many users for IP TV. In contrast, the user environment is again typically over-dimensioned. Consequently, the broadband access, being the first trusted traffic handling point in the network, is where decisions need to be enforced, and where the delay, latency and error characteristics need to be appropriate for the service needs. Applying traffic policies and controlling admission to the shared medium ensures the correct handling of all services flows and user sessions that are granted access. Premium services might take up a substantial proportion of the total resources, leading to situations in which there is insufficient headroom to accommodate peak load conditions. Also, adding a new session within the same service class might affect the quality of established sessions. Service usage must also be blocked against unauthorized or malicious users. To meet these concerns, session admission control needs to be added, that is, the network must be able to reject serv-
ice requests when the resources are insufficient or when the user is not authorized to use them. Employing admission control means that the concept of telephone networks whereby access networks maintain basic call state information is re-introduced into broadband. Within current triple-play network and service schemes, video consumes the largest bandwidth in the first mile, while the bandwidth requirements of voice are more modest. Although this will change when voice sessions evolve to multimedia, session admission control is initially of concern for IP TV and VoD. The first mile of the broadband users connection will only support a limited number of video flows (broadcast and interactive). Hence admission control will be essential in the broadband access. The ability to provide interactive multimedia services with guaranteed quality provides an opportunity for network operators to compete against providers offering these services over the best-effort Internet. It also opens the door for operators to consider replacing their revenue-generating PSTN infrastructures with pure broadband infrastructures, while maintaining the same quality and security criteria. This is in contrast to third-party providers using a best-effort Internet connection. Intelligent enforcement decisions will not be restricted to resource availability, but will also protect the network from malicious use. For example, they will protect against denial-of-service attacks on premium resources and limit the per-user signaling rate (number of messages per user per second) towards voice and video service platforms. Access and aggregation both have essential roles to play in securing the operator network from users with unfriendly intentions. In this context, broadband access equipment has an important part to play in preventing identity spoofing.
VoIP and POTS transformation

The Plain Old Telephone Service (POTS) is satisfying fewer and fewer customers. Demand is growing for home networking to support increasingly sophisticated con-
Figure 2: VoIP service delivery options
Open Services Delivery Environment

User Profile User Location Session State Softswitch
VoIP Access Gateway
PSTN
Regional Center
xDSL Residential Gateway
BRAS ISP
DSLAM/ FTTN FTTH
Video Edge IP/MPLS Internet Edge IP Routing
Video
WWW VPN
ONT
Access Network
Aggregation Network
Session Border Controller Voice
142
www.alcatel.com/atr
FIXED ACCESS VISION
sumer electronics and entertainment devices, much higher bandwidth, and carefree protection from Internet-borne security threats. The user-centric broadband vision that Alcatel shares with many service providers is that services should be useable on any device at any time. To provide an adequate response to these trends, operators are complementing their traditional voice service with VoIP with or without a multimedia component. This service will evolve to support any kind of Session Initiation Protocol (SIP) enabled terminal, whether fixed or mobile. Users expect to use the same devices whether they are at home, at the office or on the road, using a single subscription profile. As an example, the Alcatel Intelligent Mobile Redirect (IMR) solution, which is based on SIP, routes voice traffic over a broadband network or a mobile network according to the users location. It benefits users and allows operators to offer various applications (reduced charge, optimized routing, etc). IMR is the first step towards a common set of fixed/mobile solutions, which will be fully realized by the IP Multimedia Subsystem (IMS) architecture. The extension of IMS for the convergence of fixed and mobile networks is being defined in the European Telecommunications Standards Institute (ETSI) TISPAN standard. The extensions take care of PSTN replacement scenarios as well as new SIP-based VoIP and multimedia applications. Broadband access platforms can play a central role in this process. By combining narrowband with a broadband line termination on the same access platform, they provide a seamless migration from PSTN service to the new world of converged multimedia. For this purpose, VoIP gateways can be included in access equipment. In fact, VoIP access gateway technology can be deployed at various places in the network, as illustrated in Figure 2. Combining narrowband and broadband investment is an attractive option. The copper line is terminated for all services in the same equipment, which simplifies deployment and maintenance. As operators are investing in triple-play broadband delivery platforms they can, for a limited additional cost per line, use the same infrastructure to replace the aging PSTN. Narrowband linecards in broadband access equipment (copper as well as fiber) serve simultaneously as VoIP gateways. The signaling protocol can be H.248 or SIP, depending on the service profile that the operator wants to offer over the traditional phone line. Both PSTN emulation and PSTN simulation are possible. On top, combining broadband and narrowband line termination in a single broadband access equipment leads to several new opportunities for seamless service integration and service provisioning between IP-enabled terminals at the users premises and the classic black phones. The access network can play a part in correlating sessions between both technologies, and aggregating broadband and narrowband line groups towards the common network control layer.
Radio access networks have similar needs to those of copperbased broadband access. As bandwidth increases, the coverage area of the first mile radio connection decreases. For service providers that want to become active in wired as well as wireless broadband, there is an opportunity to share investments in the broadband aggregation, edge and core networks. This is particularly the case for WiMAX, which is being positioned as a radio technology to complement DSL by adding nomadicity and mobility. Incumbent Local Exchange Carriers (ILEC) can deploy WiMAX to complement DSL in remote areas. In the case of Competitive Local Exchange Carriers (CLEC), wireless technology can offer a solution for the competitive bypass of expensive landline facilities in urban and suburban areas. The second aspect of convergence is equipment sharing between fixed and mobile networks. This is already happening in the network core (IMS); it will also occur in the aggregation part and, using WiMAX, the access part. The third aspect of convergence deals with residential and business users. Fixed and mobile broadband access both rely on Ethernet, MPLS and VPLS technologies for aggregating the access nodes. It is expected that, when fiber is intensively used in the residential local loop, the same infrastructure will be used to connect business users. In a similar way to what happened with copper-connected businesses, which are by now almost all served through DSL and DSL Access Multiplexers (DSLAM), business users will want to use the same high-bandwidth capabilities for transporting their VPNs and VoIP, as well as for Internet access. Given that the whole access infrastructure will be Ethernet-based, adding business customers to this aggregation is a small step. This will be emphasized by the fact that residential users will soon be demanding as much bandwidth as business users.
Conclusion
Driven by the increased appetite for interactive broadband, broadband access networks will increasingly deploy fiber. Bandwidth and the reliability of broadband access will be tailored to cope with even the most demanding services, and investment will be aligned with service demand. Service intelligence is driving resource allocation; it will ensure that broadband investments generate money for the access network builders. Tripleplay deployments are starting now; they will be the first phase of a total network convergence that will change the way users experience the service environment. Combining fixed and mobile services and network capabilities will be next.
Convergence in Access and Aggregation

Convergence is occurring on several fronts. First, when fixed and mobile usage converge, there is a need to support user and terminal mobility on fixed broadband networks. To offer the best bandwidth at the minimum cost, various access networks will have to converge. The boundaries will blur between wireless, wireline and broadcast networks. Services crossing over different access technologies will become commonplace, with operators combining these technologies to offer innovative, inexpensive yet reliable service bundles.
Jan Van Bogaert is Strategy Director for Network Strategy in the Chief Technology Office, Antwerp, Belgium. (Jan.Van_Bogaert@alcatel.be)
Jean-Pierre Lartigue is VicePresident Marketing & Communications of Alcatel's Access Network Division, based in Antwerp, Belgium. Jean-Pierre.Lartigue@alcatel.be
Yves T'Joens is Chief Technology Officer (CTO) for the Alcatel Access Network Division, Antwerp, Belgium. yves.tjoens@alcatel.be
www.alcatel.com/atr
143
CONSOLIDATION
D. Renaudeau, D. Boettle, H. Steyaert
W IMAX: FROM FIXED WIRELESS ACCESS TO INTERNET IN THE POCKET

WiMAX access can provide new service opportunities to fixed and mobile operators thanks to its flexible radio technology and innovative features.
iMAX is at the center of the emergence of new market and technology opportunities. The widespread deployment of high-speed Internet at home has opened the door to the introduction of new services, such as video, audio, gaming and e-commerce. Today, the availability of portable devices like laptops, Personal Digital Assistants (PDA) and smartphones, is generating interest in providing similar services in nomadic conditions. In developed regions, the deployment of broadband services in remote areas is lagging behind that in (sub)urban areas. In emerging countries, low penetration of the traditional fixed telephone service and the low quality of the copper pairs is hindering large scale broadband deployment and fostering a demand for alternative wireless technologies. Broadband wireless access has been serving enterprises and operators for some years, to the great satisfaction of its users. However, the new IP-based standard developed by the Institute of Electrical and Electronics Engineers (IEEE) 802.16 and the WiMAX Forum is likely to accelerate its adoption. WiMAX will boost todays fragmented broadband wireless access market thanks to standardization and interoperability,
state-of-the-art radio efficiency with non-line-of-sight capability, and strong support from radio equipment and chipset manufacturers. By defining a generic full Internet Protocol (IP) radio access network with compelling new radio features, WiMAX can address the fixed wireless access and portable Internet markets, complementing other broadband wireless technologies (see Figure 1). This has been achieved by providing access to new spectrum resources and by offering higher bandwidth when required. WiMAX access can be easily integrated within both fixed and mobile architectures, enabling operators to integrate it within a single converged core network, thereby providing new capabilities for a user-centric broadband world.
Technology
Standardization The IEEE is facilitating the deployment of broadband wireless access by specifying the air interface for
>
Overview on IEEE802.16 standardization: http://www.ieee802.org/16/
Figure 1: Flexible WiMAX solution from wireless DSL to Internet in the pocket
802.16e Scalable OFDMA
First step to IP Mobility
Fixed Wireless Access: at Home and Office
Internet in the Pocket Nomadic User WiMAX system based on 802.16e can address both segments OFDMA: Orthogonal Frequency Division Multiple Access
wireless metropolitan area networks. The original IEEE 802.16 standard, which was published in April 2002, focused on the frequency range between 10 and 66 GHz and required line-of-sight propagation. This proved to be a key drawback which prevented cost-effective mass deployment. Consequently, a new version, which underwent several evolutions up to 802.16-2004, has been developed to expand the scope to licensed and license-exempt bands from 2 to 11 GHz. IEEE 802.16-2004 specifies the air interface, including the Media Access Control (MAC) of wireless access for fixed operation in metropolitan area networks. New market segments will be opened up by introducing handover between the radio cells. Support for this type of mobility is being defined in a working group for amendment 802.16e, the publication of which is
144
www.alcatel.com/atr
WIMAX: FROM FIXED WIRELESS ACCESS TO INTERNET IN THE POCKET
expected before the end of 2005. In general, the introduction and implementation of this version will greatly increase its attractiveness for broadband wireless access. Alcatel is strongly supporting the development of > WiMAX Forum : http://www. the 802.16e amendwimaxforum.org/home ment based on its broad experience in the field of mobile cellular system technologies. Future Alcatel products offering 802.16e features are expected to open up new market opportunities by meeting the users need for more and more nomadic communication. A key issue in the 802.16 standards is the numerous options for parameter selection. Consequently, a major objective of the WiMAX Forum is to simplify interoperability between system components by defining a limited number of profiles (sets of parameters) and to grant a WiMAX certified label. Harmonization of end-to-end architectures, including different types of core network, is also being defined in a specific Network architecture Working Group (NWG) within the WiMAX Forum. The strong presence of operators and equipment manufacturers in the WiMAX Forum is supporting the IEEE 802.16 standardization process and pushing forward the promotion and deployment of broadband wireless access in general.
SOFDMA provides additional resource allocation flexibility in the time and frequency domains (see Figure 2). Dynamically assigning the number of subchannels (SOFDMA) makes capacity allocation more flexible. Concerning the downlink and uplink capacities of WiMAX, the classical Frequency Division Duplex (FDD) mode leads to symmetrical systems. However, in the short and medium terms, asymmetric traffic is likely to be more common, making Time Division Duplex (TDD) the preferred mode. Moreover, TDD systems are more flexible, less complex and more costeffective. In the longer term, and in view of the correlated need for more extended multi-cell coverage, FDD may become more suitable. The standard allows various bandwidths to be allocated to the radio channel: 1.5 MHz (or 1.75 MHz in Europe) and multiples up to about 20 MHz. Thus the WiMAX system and the spectrum resources can be adapted to the needs of both densely and sparsely populated regions, making it suitable for use in rural and urban areas for both fixed access and portable usage. In particular, smart antenna and beam forming techniques (see Figure 3) can play a key part in enhancing performance in order to ensure compatibility with the radio sites grid of cellular operators (up to 10 km in rural areas and around 1 km in urban areas). The achievable performance in terms of distance depends on various parameters,
Innovative approach based on Figure 3: The principle of adaptive beamforming and the use of null steering to suppress 802.16e interference Alcatel considers mobility and broadband applications as basic features to be implemented in the WiMAX product. The Strong User interferer 802.16e standard provides mobility and supports the most flexible and efficient User usage of the radio channels by combinAntenna pattern X1(k) Weak adaptation with MMSE ing the advanced Orthogonal Frequency W1 interferer algorithm upon reception of a strong interferer Division Multiplexing (OFDM) scheme X2(k) Strong y(k) with the Scalable Orthogonal Frequency W2 interferer + Division Multiple Access (SOFDMA) XM(k) technique. OFDM is an optimum solution WM Antenna for robust radio wave transmission under ... ... Elements the selective fading conditions that are typical of the non-line-of-sight Adaptive Weights Beam Forming environment with multipath propagation, for very high data rates using low complexity modulators and demodulators. Subchannel-individual adaptive modulation (QPSK, 16QAM, 64QAM) increases the spectral effiincluding the environment, modulation scheme, antenna gain, ciency up to 3.75 bit/s/Hz under good propagating conditions. transmit power, antenna height and frequency band. In real non-line-of-sight conditions with a mix of modulation forFurther evolutions are planned to enhance performance, mats, the average spectral efficiency is about 2 bit/s/Hz. particularly the coverage area and bitrates. This will be achieved by implementing cost-effective Multiple Input Multiple Output (MIMO) antenna techniques adapted to WiMAX and by introducing smooth handover mechanisms for realFigure 2: Scalable OFDMA: dynamic assignment of different time services.
numbers of subcarriers to subscribers
...
Frequency [MHz] 5 2.5 0
Number of subchannels assigned to subscribers 1 4 2 8
Network Architecture
Introducing WiMAX in the network architecture WiMAX is of interest to all kinds of operators fixed and mobile, incumbent and alternative. By deploying a WiMAX access network, fixed operators will not only be able to offer wireless access, but also to offer nomadic and portable Internet services to complement their DSL and WiFi bundles.
145
10
15
20 Time [ms]
www.alcatel.com/atr
Mobile operators can integrate WiMAX into their networks to boost their portfolios by adding high bandwidth services to complement their existing data services. Access to the same applications (messaging, agenda, portal, etc) over different radio accesses is offered with a single bill and a single subscriber profile. One key requirement for WiMAX network architectures is to seamlessly integrate fixed and nomadic subscribers into the existing fixed and mobile network architectures. Alcatels WiMAX strategy is aimed at providing a comprehensive end-to-end solution tailored for all types of operator. WiMAX integration into a global seamless roaming solution for nomadic users The common services architecture approach should provide users with the following features, irrespective of the type of access network:
Figure 4: WiMAX integration into a global seamless roaming solution

WiMAX RAN WiMAX Access Control 802.1 Nomadic Portable Service Mobile RAN BSC/RNC Mobile Base Station WiMAX Base Station Multi-Access Control AAA Operator Application & Service Environment Billing Localization & Presence IP Multimedia Customer Database
Seamless mobility and roaming WiFi Hot Spots
Aggregation Network Internet AP/AC
AP/AC
WiFi/DSL at Home PSTN NGN/VoIP/SIP AC: Access Controller (WiFi) AP: Access Point (WiFi) NGN: Next Generation Network PSTN: Public Switched Telephone Network RAN: Radio Access Network RNC: Radio Network Controller SIP: Session Initiation Protocol
Common access control and BSC: Base Station Controller authentication: Users can have a single subscription and receive a single bill even though they use different access networks. The solution is based on Authentication, Authorization and Accounting (AAA) mechanisms. Access to common services: All flows are routed through the home network by using tunneling mechanisms. This interworking scenario gives the home operator full control of the service offering (including billing, policy control), and provides the user with all the services to which he or she has subscribed whatever the access network. Service continuity across different access networks through support for mobile IP. Initially users will simply want to access their usual services, but will rapidly come to demand service continuity while on the move and passing from one cell to another. The solution enables them to enjoy an always best connected experience when accessing their applications via the best available network when at home, away from home or on the move (see Figure 4).
146
Integration in a DSL network environment WiMAX enables operators to deploy wireless DSL services on their existing infrastructures: in areas where a copper infrastructure is either not available or is of poor quality; in sparsely populated areas where a wired infrastructure would not be cost-effective. As illustrated in Figure 5, connecting the WiMAX base station to the DSL Access Multiplexer (DSLAM) enables wireless DSL to be offered in the same way as Asymmetric Digital Subscriber Line (ADSL) today by reusing both the DSL access and aggregation networks as well as the existing management and provisioning platforms. WiMAX base stations can be deployed either in the central office or near remote DSLAMs, thereby making use of the infrastructure that is already owned by the operator. Collocating the base station with the DSLAM and connecting it via fast Ethernet eliminates the need for
www.alcatel.com/atr
expensive backhaul links. However, other techniques could also be used to connect the base station to the DSLAM. Internet This architecture has the advantage Gateway that existing Point-to-Point Protocol Fixed (PPP) and IP / DHCP (Dynamic Host Configuration Protocol) models can be Worldwide aggregation Nomadic Internet used to provide high-speed Internet Provider Backbone and Voice over Internet Protocol BRAS Base (VoIP) services. Owning a wireless Station Voice first-mile connection enables operGW ators to offer nomadic use, giving PSTN WiFi their subscribers the freedom to conBilling AAA DSL NMS System nect in the same way to the same servModem DSLAM ices when at home, at work or in a hot with WiMAX Access Control zone. By opting to interconnect the base station to the DSLAM, both wired and wireless, fixed and nomadic servBRAS: Broadband Remote Access Server NMS: Network Management System ices can be offered from a single platform, providing the freedom to deploy the technology best suited to a specific service and situation, The main function of the Packet Data Gateway (PDG) interthereby extending the broadband reach and the service offer. working unit is to provide a Tunnel Termination Gateway (TTG) between the WiMAX access network and the mobile core network. It also provides the following functions: chargIntegration in a mobile network environment ing gateway interfaces, IP address allocation, authentication The globally proposed strategy is to use similar mechanisms in external networks, and single access to mobile core network to those developed for integrating wireless Local Area Networks packet data domain services. These can be seen as a subset (LAN) in the mobile architecture, Third Generation Partnerof the Gateway GPRS Support Node (GGSN) functions in the ship Project (3GPP) and 3GPP2 (see Figure 6). This integra3GPP case. tion aims to provide: WiMAX access, authentication and authorization through the mobile system, IP access, meaning that as well as accessing a locally connected IP network from the wireless LAN, the user can connect through the operators core network. Evolution towards portability and Internet in the pocket The main added value of a WiMAX access network is to facilitate the evolution to a portability service for data applications or Internet in the Pocket, making it a natural complement to mobile and WiFi networks. Portability will offer some simple mobility functions through handover capabilities between access points, while delivering the same user experience as fixed access or nomadic use when the user is stationary. Handover will be based on IEEE 802.16e and best-effort mechanisms, leading to simpler network and user device architectures and software. All delay-tolerant applications will be able to resume IP connections or reconnect on notification of a new pointof-attachment. This portability service does not presume a particular range of mobility speeds, but by relaxing the constraints on handover latency it makes this mode of operation more suitable for reduced mobility rates and for supporting IP services that can tolerate some performance degradation during handovers (non-real-time services). This portability function relies on handover mechanisms embedded in the 802.16e air interface, on mobile IP principles and on the introduction of home agents in the network (see Figure 7). Both micro-mobility (handover between WiMAX base stations managed by the same access control) and macro-mobility (handover between WiMAX base stations managed by different access controls) are supported. Furthermore, inter-technology mobility functions enable the access network to change (between WiMAX, WiFi, and Mobile 2G or 3G networks) while an application session is ongoing. This feature is the cornerstone of the integration of multiple radio access networks with a single converged core network.
147
Figure 5: WiMAX integrated in a DSL architecture
Figure 6: WiMAX integrated in a mobile architecture
Direct Internet Access
WiMAX Access Network
Mobile AAA Server Packet Data Gateway
Mobile Operator Architecture
Mobile PS Services including Internet Access
Such an approach enables the user to access operator services and have identical access and authorization mechanisms based on a Subscriber Identification Module (SIM) and Universal SIM (USIM).
www.alcatel.com/atr
Radio Access Network

Challenges Operators deploying radio access networks for broadband services face several challenges that WiMAX radio infrastructure suppliers should answer. Two main challenges are: Minimizing the operators investment by reducing the number of radio sites. Containing costs for backhauling.
Figure 7: Mobility in WiMAX
Mobile RAN
BSC/RNC
WiMAX RAN
WiMAX Access Control 3GPP/3GPP2 Core Network
Reduced number of radio sites Overall, the WiMAX network deployment strategy is to deliver broadband nomadic services with a large coverage. When deploying any new radio technology and new services, minimizing the required capital expenditure is a key factor for successful implementation as it allows the operator to achieve its financial targets. Because the major part of an operators network investment is in installing the radio sites, it is essential to achieve the required coverage with the minimum number of sites. As environmental concerns are leading to sites becoming a scare resource, an appropriate strategy would be to reuse the existing cellular network grid and to favor solutions for sharing sites between existing cellular network sites and WiMAX networks. Compared with existing cellular systems, WiMAX systems implement advanced radio features that compensate for the extra attenuation resulting from: Higher carrier frequency: the first WiMAX deployments will be in the 2.5 and 3.5 GHz bands. Larger transmission bandwidth: a typical WiMAX bandwidth is 10 MHz, compared with 200 kHz for GSM systems and 5 MHz for wideband code division multiple access systems. Deep indoor penetration: indoor coverage is mandatory for WiMAX in any environment, whereas with cellular systems it is only essential in urban areas. The IEEE 802.16 standards define advanced radio features, including smart antenna technologies and uplink subchanneling to realize the desired performance, that is, high data rates of several Mbit/s in a cellular environment.
Home Agent WiMAX Access Control AP/AC AP/AC
WiFi Hot Spots
Backhauling Introducing WiMAX at radio sites makes it necessary to backhaul the traffic generated at this point (around 20 Mbit/s per cell), as shown in Figure 8. Thus the launch of nomadic broadband services raises new technological challenges, including questions related to the capacity and cost of different backhauling methods and technologies. Alcatels WiMAX systems are based on an all-IP network architecture relying on IP/Ethernet transport. The cost advantages of this approach are considerable; converging voice and data onto a single network reduces both the initial capital expenditure and the operating expenses. WiMAX solutions should include radio site backhauling options for different kinds of physical interfaces, mainly a wireless solution (e.g. point-to-point or WiMAX self-backhauling), a copper solution (e.g. bonding, that is, combining symmetric high speed DSL copper pairs, could offer the required throughput) and a fiber solution.
Conclusion
The latest developments of the IEEE 802.16 standard based on revision e together with the end-to-end standardization initiatives now integrated in the WiMAX Forum through the Network Working Group are driving the emergence of innovative broadband wireless access solutions. At the same time, the success of the WiMAX Forum, with more than 200 members and backed by industry leaders, including operators, component suppliers and infrastructure suppliers, is confirming the promising future of WiMAX as an access technology. The combination of both advanced radio features and flexible end-to-end architectures makes WiMAX an attractive solution for diverse operators: fixed (including DSL), mobile and wireless Internet service providers. It can help to bridge the digital divide and make portable Internet a reality with
Figure 8: Backhauling
IP/Ethernet
WiMAX Backhaul How ?
WiMAX Base Station
Access Control
148
www.alcatel.com/atr
WiMAX chipsets being embedded in laptops, PDAs and other portable devices. Integrated mobility and roaming solutions between various accesses will enable users to enjoy an always best connected experience.
David Renaudeau heads the Product Management department of the Broadband Wireless Access Business Unit in the Mobile Radio Division of the Alcatel Mobile Communications Group, Vlizy, France. (David.Renaudeau@alcatel.fr)
Hendrik Steyaert is Portfolio Manager for wireless access and home networking technologies in the Chief Technology Office of the Alcatel Access Networks Division, Antwerp, Belgium. (Hendrik.Steyaert@alcatel.be)
Dietrich Boettle is head of the Access Network Department and Project Leader for the Alcatel Wireless DSL/WiMAX project in the Research and Innovation department, Stuttgart, Germany. (Dietrich.Boettle@alcatel.de)
www.alcatel.com/atr
149
CONSOLIDATION
F. Engmann, G. Walz
ALCATEL EVOLIUM MULTI-STANDARD RADIO ACCESS SYSTEM

Multi-standard radio access is the new trend in network design! So what are the requirements and benefits?
n 1998, Alcatel launched the Evolium concept. Initially it was implemented for the base station and the operations and maintenance system, but it has now been extended to encompass the entire radio and core network product range. Alcatels development strategy rests on three key principles: Continuous re-engineering and cost optimization utilizing the latest technologies. Long-term solution offering maximum upstream and downstream compatibility. Multi-standard radio access solution. The multi-standard radio access solution, which combines several radio technologies in one system, is an important strategic principle for Alcatel and will govern the future evolution of its mobile communication networks. Mobile operators today face a major challenge: how to efficiently introduce new technologies, like Enhanced Data rates for GSM Evolution (EDGE), Universal Mobile Telecommunications System (UMTS), High Speed Downlink Packet Access (HSDPA), Multimedia Broadcast / Multicast Service (MBMS), the WiMAX system (Worldwide Interoperability for Microwave Access) and transmission systems based on the Internet Protocol (IP), into their existing mobile networks.
System for Mobile communication (GSM) or UMTS. Mobile communication is evolving towards multimedia services: This requires the management of multiple Quality of Service (QoS) classes to provide optimized support for new services like video and audio streaming and multicast applications. Wireless broadband access: Innovative concepts and technologies are driving the radio interface to higher bandwidth efficiencies, providing higher throughputs and lower costs for wireless data transmission. Rapidly changing mobile communication market: The current mobile market is being strongly influenced by the move from second to third generation technology, by the implementation of new applications and services, and by the emergence of new competitors and radio technologies, such as Public Wireless Local Area Networks (PWLAN), WiMAX and UMTS Time Division Duplexing (TDD). Increasingly competitive mobile market: In the established GSM markets, equipment is rapidly becoming a commodity. This kind of market with its strong competition and a clear need for differentiation has a significant impact on equipment design and network implementation. Analyzing these market requirements in more detail and filtering out the main future requirements and evolutionary paths leads to two main convergence scenarios: Service convergence The focus of this convergence is to meet the communication needs of mobile subscribers regardless of the radio access technologies. Ideally, the user should not have to take account of the various technologies, and need not know whether he or she is currently using GSM, UMTS or another access technology.
Key Market Drivers for Multi-standard Radio Access

Irrespective of their chosen network evolution path, mobile operators face one significant business challenge: attracting subscribers by seamlessly introducing a large number of compelling new services that satisfy users needs and combine high quality with acceptable prices. This needs to be achieved while optimizing network investment. The following main market trends are driving evolution: User-centric service offering: Today, network operators are planning to offer a rage of communication services, not simply access to a specific radio technology, such as the Global
150
www.alcatel.com/atr
The aim is to support the same user services over different radio access technologies. The only differences in the offered services would be the result of different radio spectrum efficiencies and the available bandwidth of the radio interfaces. Alcatel strongly believes that service convergence will lead to the convergence of infrastructure technologies.
Figure 1: Functional split of Evolium Node B implementation
SUMU module
Baseband module
TEU module TX PA
ANRU module Antenna Network (Duplexer)
O&M Transmission
Baseband Processing RX LNA
Ant' TX/RX
Node B cabinet
Platform convergence PA: Power Amplifier In line with its strategy of reducing the operators operating costs and minimizing their capital expenditure, Alcatel is working on the convergence of the various radio access platforms, such as GSM, UMTS and WiMAX. The following are the main drivers behind platform convergence:
LNA: Low Noise Amplifier
TX/RX: Transmitter / Receiver
resolve the paradox currently facing the mobile industry: how to efficiently implement new technologies while optimizing the use of earlier investment and ensuring risk-free network migration. How does the Evolium BTS meet these design requirements? The following are the cornerstones of the BTS design. Single equipment practice Alcatel uses the same equipment practice for all its rack systems, enabling all the available modules to be used in all versions of the multi-standard base station (e.g. indoor, outdoor and compact versions). Future developments aimed at optimizing the rack system will take account of this equipment practice and will provide backward compatibility. The use of a common equipment practice is fundamental to providing flexibility and scalability over the whole product range. It enables modules using different radio technologies to be freely mixed in the same rack (e.g. GSM, UMTS and, in the future, WiMAX). In addition, a multi-standard rack solution will make it possible to reuse existing foundations, structures, feeders (in most cases) and masts. No change of building permit is required and no additional construction effort is necessary, making it possible to deploy UMTS, for example, very rapidly. The fully flexible use of the internal rack space for GSM and UMTS modules allows the operator to adapt its network smoothly as market requirements change. For example, where UMTS is successful the capacity can easily be increased while GSM modules are taken out of service as demand declines. Modular design A key feature of the Alcatel multi-standard base station design is its modularity. Figure 1 shows the functional elements of a UMTS Node B (black text) and how Alcatel has implemented them in only four types of module (colored): SUMU: Station Unit Module, which implements the BTS control functions. Baseband Module: performs the baseband / signal processing functions (Figure 2). TEU: Transmission Unit Module + power amplifier (multi carrier). ANRU: Reception Unit Module + antenna network (function duplexer). The same approach is used for the GSM part, which uses just three standard modules: the transmitter / receiver module, the
151
Reduction in operating expenses (e.g. for transmission lines and site engineering). Cost savings in network implementation and maintenance. Common network management principles for all radio technologies, enabling operations and maintenance staff to be used efficiently. Seamless roaming between the different radio technologies without any service interruption. Optimization of radio bandwidth usage by service and application differentiation on the various radio technologies (e.g. high speed data on UMTS and voice on GSM). The first fruits of this convergence policy are Alcatels multistandard Base Transceiver Station (BTS), which handles the different radio technologies (e.g. GSM, UMTS/HSDPA and WiMAX) on a single platform, and the associated multi-standard management platform and tool chain.
Multi-standard BTS Ready for Future Technologies

Based on market requirements, the design of the multi-radio access system should meet the following main criteria: Flexibility to cope with uncertain market conditions, enabling operators to react rapidly to new service and technology requirements. Scalability to allow incremental investment in new technologies. Full backward compatibility to maximize the reuse of previous investment. Rapid introduction of new technologies without the need for hardware changes. Reduced operating costs and minimum capital expenditure. Meeting these design criteria is the only way to
www.alcatel.com/atr
antenna network, which implements the combiner and duplexer functions, and the SUMU, which provides the BTS control functions. The transmitter / receiver module performs digital baseband processing for GSM, eliminating the need for a separate baseband module. All the GSM and UMTS modules follow the same design guidelines and equipment practice, as will modules for future radio technologies (e.g. WiMAX). Consequently, a very flexible implementation with a free mix of modules is possible within the same rack system according to the operators needs. Hardware implementation based on digital processors The designs of all Evolium multi-standard BTS modules follow one strict rule: the use of programmable digital technology where it is economically justified. Thus all modules use microprocessor technology, in most cases Digital Signaling Processors (DSP) which can be loaded with the appropriate firmware and application software. For example, the baseband board shown in Figure 2 is a freely programmable digital processing resource. This approach allows new features to be introduced into the network simply by changing the software without replacing any hardware. Good examples are the network upgrades for the introduction of EDGE and HSDPA, which only required a software upgrade to implement these important services in the mobile network. Evolium BTS design: smooth system evolution A fundamental advantage of the Evolium BTS design is the simple, smooth introduction of new radio technologies, for example, to follow the Third Generation Partnership Project (3GPP) UMTS system evolution, as shown in Figure 3. In principle, this can be achieved in two ways: Implementation of new features simply by upgrading the software.
Figure 3: 3GPP UMTS system evolution
Figure 2: Digital baseband processing board
Introduction of new modules into the existing rack system with full backward compatibility.
Software-only upgrades were required in the past to introduce the General Packet Radio Service (GPRS) and EDGE, and will be used to deploy High Speed Data Packet Access in the downlink and uplink directions (HSDPA and HSUPA, respectively), as well as smart antenna concepts. However, in some cases additional Smart antenna modules will be needed, for example, An antenna system to support Multiple Input / Multiple that combines Output (MIMO) techniques and new multiple antenna frequency bands for UMTS. These elements with a modules use the same equipment pracsignal-processing tice, enabling them to be deployed in capability to the existing rack system and operate optimize its alongside existing modules (e.g. GSM radiation and/or and UMTS). Full backward compatibilreception pattern ity is guaranteed. The only limitation automatically in is the space available within the rack. response to the This is also true for the modules signal environment. that will be developed to support new radio technologies like Orthogonal Frequency Division Multiplex (OFDM) and WiMAX. These new radio modules can be installed in existing GSM and UMTS multi-standard BTS. One additional challenge for mobile network operators will be the introduction of IP transport technology into their existing GSM and UMTS networks within the next few years. Alcatels digital processor approach will enable IP to be introduced simply by upgrading the software.
Step 1:
Cell User Peak rate up to 10.8 Mbit/s > Adaptative Modulation & Coding Scheme (QPSK & 16QAM) > MAC Packet Scheduler with ARQ in Node B
Multi-standard Operations and Maintenance

3GPP R5 3GPP R6 3GPP R7 RAN Evolution
Step 2:
Users numbers with Higher throughput per cell, Coverage Extension Balance with UL: 2Mbit/s per cell > Smart Antennas > HSUPA
Step 3a: Evolution

Capacity/Cell User throughput enhancements > MIMO
Step 3b: Evolution

Capacity/Cell User throughput enhancements > OFDM
ARQ: Automatic Request for retransmission MAC: Media Access Control QAM: Quadrature Amplitude Modulation
QPSK: Quadrature Phase Shift Keying RAN: Radio Access Network
Operations and Maintenance (O&M) has a significant impact on the operators operating expenses. Important cost factors include not only network maintenance and spare parts management, but also network supervision and fault management, which require staff to be available 24 hours a day throughout the year. The deployment of multi-standard networks will make it necessary to identify synergies in the area of network management and maintenance. While maintenance procedures and spare parts handling are implicitly simplified by the Evolium multi-
152
www.alcatel.com/atr
standard BTS practice, efficient network management of multi-standard radio access networks requires some additional efforts on the management system side. Multi-standard O&M functions By collating the needs of various operators, Alcatel has identified the areas where synergies in the management of different radio access technologies (in particular 2G and 3G, but also technologies like WiMAX) are of the utmost importance and justify the development of multi-standard O&M functions: Fault management: The process of fault detection and tracking is the same whatever the core technology of the mobile network. The supervisors job is the same, and staff must be available round the clock. Common QoS monitoring: An operators various radio access networks interact, making it necessary to assess the influence of one network on the others (e.g. interference, blocking, handover, load balancing). Thus, a common analysis of the QoS of all the affected radio access networks is essential in order to optimize the multi-standard network. Cross adjacencies declaration: The operator of a multi-standard network should have mechanisms available for optimizing resource management in the different radio access technologies, in particular for 2G and 3G networks. Handover from one technology to another is an essential resource allocation mechanism, which is based on the load and coverage conditions in the network or on the types of service. Declaration of adjacencies between cells supported by different radio technologies (cross-adjacency) will ease the management of those mechanisms. Platform administration and security: The advent of new radio access technologies will make it necessary for mobile operators to renew their information systems to manage several types of network simultaneously. Special attention should be paid to user management (security) and hardware
administration. A common repository of users and their access rights would be particularly useful. Common administration procedures can also be beneficial. Evolium multi-standard portal To cope with these fundamental requirements, Alcatel has introduced the concept of the Evolium multistandard portal (see Figure 4), which will provide the required multi-standard O&M functions on top of existing 2G and 3G O&M features and Operations and Maintenance Centers Radio (OMC-R). At the same time, the portal is prepared for the introduction of forthcoming radio access technologies, such as WiMAX. The multi-standard portal provides all the O&M functions for which a multi-standard approach is reasonable; it communicates with the networks via the OMC-Rs, which also provide the technologyspecific functions. The multi-standard portal and OMC-R functions can be accessed through a unique user terminal at the portal level. A major feature of the multi-standard portal is its network supervision functions:
Orthogonal Frequency Division Multiplexing (OFDM) A modulation technique that splits a data signal into multiple lower bit rate signals, which are transmitted simultaneously on different subcarriers around a central frequency to the receiver. This kind of modulation is claimed to be robust in multipath environment. Modulation used in particular in ETSI Hiperlan 2 and IEEE 802.11a/g.
Common window for alarms. Common QoS viewer. Common supervision tree and map. Common security management features. Navigation to native user interfaces for non multi-standard actions (radio configuration, network element declaration, etc). The multi-standard portal approach is complemented by features that support network analysis, stability monitoring and optimization tasks. These features are built around a data warehouse combining all the sources of network data that are relevant for the daily network analysis and optimization, allowing complex composite indicators to be computed. Multi-standard O&M without network constraints The multi-standard portal concept allows the various radio access technology product lines to evolve independently because each technology has its dedicated OMC-R. The portal in no way constrains the network. It can accommodate the different migration paths of the single-standard OMC-Rs of the underlying radio access network parts. An operator that has not yet deployed any of the radio access technologies (2G, 3G, WiMAX, etc) will not be affected by the evolution of these technologies.
Figure 4: Multi-standard portal concept
Multi-standard PORTAL
2G OMC-R
3G OMC-R
Conclusion
2G NE
MBS: Multi-standard BTS
MBS
3G NE
NE: Network Element
Alcatel has been supplying mobile radio technologies since 1991, starting with GSM. This early experience has helped Alcatel to build a long-term vision for its product line, resulting in the Evolium concept. Since then its success has continued to grow along with Alcatels market share.
153
www.alcatel.com/atr
The Evolium multi-standard concept has been well received by the market because of its flexible and scalable system design, which provides benefits for the operator in terms of capital expenditure and operating expenses. In particular, the platform convergence principle, with its future-proof concept of introducing new technologies like EDGE and HSDPA simply by upgrading the software, and implementing new radio standards (e.g. UMTS) by adding new modules, has received considerable attention from mobile operators. The operations and maintenance concept fully supports the Evolium multi-standard solution through the multi-standard portal. It enables common management principles to be applied to all the supported radio technologies, thereby benefiting from the efficient allocation of O&M staff and reduced training effort. Because of its success, Alcatel will continue to develop the multi-standard Evolium concept in order to support more features and radio technologies on one common platform.
Frank Engmann is Product Manager in Business support GSM/UMTS, in the Mobile Radio Access Business Division of the Alcatel Mobile Communications Group, Stuttgart, Germany. (FEngmann@alcatel.de)
Gunther Walz is Product Manager, GSM/UMTS Operations and Maintenance, in the Mobile Radio Access Business Division of the Alcatel Mobile Communications Group, Stuttgart, Germany. (G.Walz@alcatel.de)
154
www.alcatel.com/atr
TECHNICAL PAPER
OPERATIONAL EXCELLENCE
T. Bostoen, R. Oehen, J. Verlinden
OPTIMIZING DSL FOR MULTIMEDIA SERVICES

Line optimization can increase DSL speed, quality and stability while reducing operating expenses.
n the initial stages of Digital Subscriber Line (DSL) deployment, operators were quite conservative regarding their DSL service offerings. First, the offered data rates were low, which meant that the majority of lines had a high noise margin. Second, high-speed Internet was the only DSL service, and was quite robust against quality and stability problems at the DSL physical layer. Thanks to this conservative service offering, operating expenses were well contained. To maximize the coverage of new multimedia services, lines must be operated at the highest possible data rate while increasing both the quality and stability. However, increasing the data rate reduces the noise margin. A surge in quality and stability problems might cause an increase in help desk calls, truck rolls and churn, leading to increased operating expenses. This risk can be ameliorated by raising the level of intelligence and automation in DSL loop qualification, in-service optimization and problem diagnosis. First, loop qualification can be improved by exploiting the line testing capabilities of the DSL transceiver. Second, when a line suffers from performance degradation or fails to synchronize, the root cause can be rapidly diagnosed by inspecting the operational data acquired in-service or using the line testing functionality of the DSL chipset. Thirdly, the configuration parameters of lines that are in-service can be optimized automatically according to the line conditions. In short, DSL becomes self-managed, as explained in the first part of the article. While DSL self-management helps to maximize DSL rate/reach performance in the short term, advances in DSL technology are expected to improve performance significantly in the long term, as explained
in the second part of the article on everfaster DSL, which focuses on techniques that avoid or mitigate crosstalk: Dynamic Spectrum Management (DSM) and Multiple-Input Multiple-Output (MIMO) transmission.
Self-Managed DSL
Line quality and stability Quality and stability are gaining in importance with the introduction of new multimedia services that are less tolerant to transmission errors and service interruptions. At the DSL physical layer, quality is defined as the Mean Time Between Error events (MTBE) and stability as the Mean Time Between (spontaneous) Resynchronizations (MTBR): MTBE (measured in bits) expresses the average number of bits between two error events, for example, a code violation or Cyclic Redundancy Check (CRC) error. MTBR (measured in seconds) measures the average time between two spontaneous resynchronizations, that is, line re-initializations not induced by the user. The values of MTBE and MTBR that guarantee sufficient quality and stability are service dependent. For example, MTBE and MTBR have to be higher for video than for high-speed Internet. Resynchronizations and code violations can have various causes, such as: Stepwise crosstalk: This is the change in stationary noise levels when modems on neighboring lines are switched on and off. Impulse noise occurs at random. It is non-stationary noise from voltage spikes generated on the line when electricity is switched on and off, including light switches, electric motors (e.g. refrigerators, air conditioning units) and home appliances. Micro cuts in the copper wire and intermittent contacts lead to random micro-interruptions of the DSL signal. DSL service lifecycle DSL self-management helps to ensure quality and stability when lines are pushed to their speed limits to support new multimedia services. This is achieved by raising the level of intelligence and automation in the management of DSL lines during their service lifecycle, which includes the following stages (see Figure 1): Order: Service order generation. Loop qualification: Checks whether a service can be delivered.
155
Dynamic Spectrum Management (DSM) Dynamic Spectrum Management (DSM) aims at avoiding or mitigating crosstalk between DSL lines in the same binder with the goal of improving the rate/reach performance of these Digital Subscriber Lines. Multiple-Output Multiple Input (MIMO) Multi-user transmission: the signals of multiple DSL lines are jointly processed in order to mitigate crosstalk interference.
www.alcatel.com/atr
Figure 1: DSL service lifecycle

Discontinuity (e.g service upgrade)
Loop qualification by means of: SELT (before CPE is installed) DELT (for service upgrade)
Validation Provisioning Activation
Operations in-service optimization by means of DLM Problem diagnosis by means of: SELT (line cannot get into service or into loop diagnostics mode) DELT (only degraded performance)
Provisioning: Physical and logical allocation of resources. Activation: service activation. Validation is not currently included in the DSL service lifecycle of most operators. It is proposed to introduce this stage to verify that a line operates with sufficient quality and stability for the service it runs. Verification is impossible at the loop qualification stage, because of the unpredictable nature of the causes of quality and stability problems. Operations: normal service operation. Decommissioning: De-provisioning the service resources. Loop qualification The introduction of line testing capabilities in the DSL transceivers can significantly improve the reliability of the loop qualification process. A distinction can be made between DualEnded Line Testing (DELT) and SingleEnded Line Testing (SELT). DELT is dual-ended in the sense that it requires both the Central Office / Remote Terminal (CO/RT) and Customer Premises Equipment (CPE) DSL transceivers to be connected to the line and powered on. SELT is single-ended as it is used when CPE is not installed. Consequently, line testing is performed by the CO/RT DSL transceiver alone. Both SELT and DELT are based on measurements performed by the DSL transceiver. Typically, an expert system interprets the measurements. This expert system, which is located on a server in the operators intranet, retrieves the measurement data via the management information base in the DSL Access Multiplexer (DSLAM). Increasing loop qualification accuracy will reduce the number of false positives and false negatives. A false negative occurs when a line is disqualified for a particular DSL service although it is capable of supporting the service. This leads to lost revenue. Even worse is a
156
Order
false positive, which means that a line is qualified for a DSL service even though it cannot support that service. This results in unnecessary help desk calls and truck rolls to resolve the rate, quality or stability problem. The customer becomes frustrated, possibly leading to churn. As false positives are worse than false negatives, operators have avoided the former by offering overly conservative DSL services and using too stringent deployment rules, which also lead to lost revenue. Improved line qualification thus helps an operator to expand its DSL coverage and allows lines to be upgraded to higher data rates. As SELT and DELT measurements are performed by the DSL transceiver, no additional hardware, such as hand-held testers, integrated test unit boards, test access matrices, and external test heads, is required. Two components are required for SELT: the measurement capability of the
DSL transceiver, called the SELT-PMD (Physical Medium Dependent) layer, and a processing unit, called the SELT-P (Processing), which is typically an expert system. The SELT-PMD layer is based on reflectometry. An excitation signal is inserted in the loop by the CO/RT DSL transceiver, which also measures the echo response. SELT-P processes the echo response to derive, for example, the loop length, attenuation and attainable bitrate. Currently SELT is being standard-
>
Single-Ended Line Testing (SELT), Alcatel White Paper, to be published 3rd Quarter 2005 on www.alcatel.com.
Figure 2: Example of in-service performance monitoring

Contour plot Bit Loading
www.alcatel.com/atr
ized in the International Telecommunications Unions (ITU) G.selt project. DELT has been standardized in ADSL2 as a special loop diagnostic mode that enables measurements to be made even when line conditions are too poor for the line to go into service. It enables the loop attenuation to be measured directly. For example, to measure the downstream attenuation, an excitation signal is injected into the loop by the DSL transceiver in the central office and the frequency response is measured by the DSL transceiver in the CPE. If the line does come into service, then the same DELT measurements as collected in the loop diagnostics mode can be gathered by ADSL2 in-service monitoring, which is already available in ADSL1, although limited with respect to the amount of data acquired. For example, in ADSL1 only the bit loading is measured per subcarrier (see Figure 2), whereas in ADSL2 the channel transfer function, quiet line noise (noise on the line when there is no DSL signal) and signal-to-noise ratio are also measured per subcarrier.
Each time the profile changes, the line resynchronizes to apply the new configuration parameters. DLM enhances network integrity, especially when the DSL network is undergoing significant infrastructure changes, such as rate upgrades, introduction of new DSL technologies and remote terminals. It reduces call center costs, network operations and field intervention costs, and churn. DLM works as follows. First, a DLM service template (see Figure 3) is defined, consisting of a set of line configuration profiles and MTBE and MTBR thresholds for a given service. A DSL line is assigned to a particular DLM service template according to the service it has to support. The line is initialized with one of the profiles given by the DLM service template. Next the MTBE and MTBR are monitored in-service and compared with the thresholds defined in the DLM service template. If one of the neighboring profiles in the profile hierarchy is better suited to this line, then the line is reinitialized with the new profile. This pro-
Quiet Line Noise (QLN) Noise present on the line, when it doesnt carry DSL signals. Interleaver depth The number of codewords stored in the memory of the interleaver. The larger the interleaver depth, the better the protection of the DSL line against noise bursts.
depth is called Dynamic Code Management (DCM). Stabilize: Switch to a profile that ensures a higher MTBE and MTBR. This can be achieved by either lowering the bit rate (DRM), increasing the interleaver depth (DCM), or increasing the transmit Power Spectral Density (PSD), which is called Dynamic Spectrum Management (DSM). Reduce crosstalk generated into neighboring lines: Switch to a profile with reduced transmit PSD (DSM), which means accepting a smaller MTBE and MTBR. This may result in an opportunity to upgrade neighboring lines. More information on DLM can be found in [1]. It is important to note that not all loop impairments can be mitigated by DLM. In many cases, field intervention will be necessary to repair the line. Problem diagnosis When the performance of a line is degraded or it can no longer synchronize, it is important to detect and locate the fault as soon as possible. The root cause can be identified by using an expert system to interpret the DELT and SELT measurement data. Fast reliable diagnosis of the faulty line, even before a user complains, reduces the number of help desk calls and field interventions required for the line repair. When repairs are performed proactively, users should not be aware there was a problem, thereby helping to reduce churn. Figure 4 shows that the DSL copper loop is a challenging transmission environment requiring intelligent, automated and proactive fault detection and location. Self-managed DSL is supported by the Alcatel 5530 Network Analyzer [2] in combination with the Alcatel 730x ASAM/ISAM (Advanced Services Access Manager / Intelligent Services Access Multiplexer).
Figure 3: DLM service template

Dynamic Rate Management (DRM) 4 Mbit/s Fast -40 Bm/Hz 4 Mbit/s Fast -46 Bm/Hz 4 Mbit/s Fast -52 Bm/Hz Dynamic Spectrum Management (DSM) 2 Mbit/s Fast -40 Bm/Hz 2 Mbit/s Interleaved -40 Bm/Hz Profile Upgrade 1 Mbit/s Fast -40 Bm/Hz 1 Mbit/s Interleaved -40 Bm/Hz Stabilize Reduce crosstalk 3 Mbit/s Fast -40 Bm/Hz 3 Mbit/s Interleaved -40 Bm/Hz 4 Mbit/s Interleaved -40 Bm/Hz Dynamic Code Management (DCM)
In-service optimization Once the line is in service, the line configuration parameters can be optimized according to the line conditions in order to achieve the highest possible data rate while ensuring adequate quality and stability. Optimization can be fully automated by using Dynamic Line Management (DLM), which automatically assigns the profile best suited for a particular line supporting a given service.
cedure is repeated until the line has converged to the optimal profile. DLM switches a line to a different profile in an attempt to: Upgrade: Switch to a profile that offers a higher bitrate or shorter latency, which means accepting a smaller MTBE and MTBR. Adjusting the bitrate is called Dynamic Rate Management (DRM), while adjusting the interleaver
Ever-faster DSL
While DSL self-management helps to maximize DSL rate/reach performance in
157
www.alcatel.com/atr
the short term, advances in DSL technology are expected to improve performance significantly in the longer term. In principle there are six ways to increase the capacity of DSL lines: Improve DSL transmission technology: For example, ADSL2 adds an average of 50 to 80 kbit/s of capacity to a typical line compared with ADSL. Increase DSL transmission bandwidth: ADSL2+, for example, doubles the transmission bandwidth from 1.1 MHz in ADSL to 2.2 MHz, thereby optimizing capacity on short lines. Increase DSL transmit power and/or PSD: READSL2, for instance, operates with a boosted PSD mask to increase coverage. Reduce subscriber loop attenuation by moving to new fiber deployment scenarios (Fiber to the x; FTTx). Reduce noise level received by the DSL modem, for example, by rewiring the DSLAM port for the repeatered DSL at the MDF to another pair, possibly in a different binder. Bonding. For example, if there is a quad instead of a pair running to every subscriber, the capacity can be doubled by transmitting over both wire pairs. The rest of the article focuses on reducing the noise level, and in particular on techniques that help to avoid or mitigate crosstalk between pairs in the same binder.
Figure 4: Copper loop impairments
bridged tap
Central office
DSLAM MDF cable cut Feeder crosstalk
RFI
impulse noise POTS/ISDN FDI distribution CPE STB untwisted drop wire
wiring fault
short degraded contact FDI load coil split
short between pairs
resplit crosstalk
untwisted in-house wires
POTS/ISDN CPE
fiber cut RT short to ground
missing splitter
in house bridged tap
FDI: Feeder Distribution Interface ISDN: Integrated Services Digital Network MDF: Main Distribution Frame POTS: Plain Old Telephone Service
RFI: Radio Frequency Interference RT: Remote Terminal STB: Set-Top Box
Dynamic spectrum management In [3], a distinction is made between DSM at levels 0, 1, 2 and 3, according to the degree of coordination between multiple DSL lines: Level 0: Static Spectrum Management (SSM). A DSL line maximizes its own performance without considering the performance of neighboring lines. Spectral compatibility between lines in the same binder is ensured by imposing restrictions on the transmit power and spectrum.
158
Level 1: Autonomous (single-user) power allocation. The power on a DSL line is allocated so as to avoid unnecessary crosstalk to its neighbors. This increases the capacity of the binder, if applied by all or most of the lines in the binder. At level 1, the power allocation of a DSL line is computed based on its own condition and service requirements, without any coordination with other lines in the same binder. Level 2: Coordinated (multi-user) power allocation. Like level 1 it aims to avoid crosstalk by adapting the transmit spectra, but power allocation is based not only on its own line condition and service requirements, but also on those of other lines. This requires coordination between lines in the same binder. DSM level 2 ultimately allows the optimal spectrum allocation for each line in the binder to be computed, thus maximizing the total binder capacity. Level 3: Multi-user transmission aimed at mitigating crosstalk. This is actually MIMO transmission, as discussed below.
DSM levels 1 and 2 can significantly improve performance in scenarios where central office lines and remote terminal lines are present in the same binder. More details can be found in [4]. Multiple-Input Multiple-Output transmission The objective of MIMO transmission is to mitigate crosstalk by jointly processing the signals of multiple lines in a binder. To make joint processing possible, either all transmitters and/or all receivers must be collocated. The crosstalk mitigation method will depend on the collocation, as shown in Table 1. MIMO can be used both for point-topoint connections and for point-to-multipoint connections. In the former case, all processing can be performed by the receivers, while in the latter case (e.g. one central office with multiple CPEs), all processing is carried out at the central office. Using MIMO, the binder is considered as a whole, instead of looking at each line individually. Hence the term MIMO (Multiple-Input Multiple-Output).
Table 1: Crosstalk mitigation method according to transmitter and receiver collocation

Collocated Receivers Collocated Transmitters Non-Collocated Transmitters Crosstalk Cancellation Crosstalk Cancellation Non-Collocated Receivers Crosstalk Precompensation MIMO not possible
www.alcatel.com/atr
MIMO is relatively complex as all the lines in a binder have to be processed by a single entity. Thus either the processing of all lines has to be performed on one chip, or fast data connections (to exchange crosstalk information) are needed between all the chips that process the lines. Nevertheless, MIMO has the potential to significantly improve the performance on short loops, as shown by Figure 5 [4]. DSM and MIMO transmission are still in the research phase, although a basic implementation of DSM is supported by the Alcatel 5530 Network Analyzer [2], as explained earlier.
Figure 5: MIMO transmission performance gains for VDSL (yellow) without MIMO, and (green) with MIMO
DSM gain [m] Bandwidth tier (Mbit/s) MIMO Upstream performance 50 25 10 5 3 1 Downstream performance (AWGN:-140dBm/Hz) 50 25 10 5 Downstream performance (ANSI-A alien noise environment) 50 25 10 5 0 0.5 1 Reach [km] 1.5 2 +620 +160 +30 +10 +720 +280 +190 +0 +330 +620 +300 +140 +100 +30
Conclusion
The concept of self-managed DSL [5] proposes to increase intelligence and automation in loop qualification, in-service optimization and problem diagnosis. This new approach to DSL management will help operators to optimize their DSL lines for multimedia services that require higher quality and stability. As a consequence, DSL self-management can minimize help desk calls, field interventions and churn, thereby reducing operating expenses. It is based on the intelligent interpretation of operational, DELT and SELT measurement data together with dynamic line management, which automatically optimizes the configuration parameters in-service. Self-managed DSL is supported by the Alcatel 5530 Network Analyzer [2] in combination with the Alcatel 730x ASAM/ISAM.
Longer-term evolution of DSL technology will lead to even higher speeds. Significant gains in rate/reach performance are expected thanks to advances in techniques such as dynamic spectrum management and MIMO transmission.
Tom Bostoen is Product Manager for the Alcatel 5530 Network Analyzer in the Alcatel Networks Division, Antwerp, Belgium. (Tom.Bostoen@alcatel.be)
References
[1] Dynamic Line Management for Digital Subscriber Lines, Alcatel White Paper, to be published 3rd Quarter 2005 on www.alcatel.com. [2] Alcatel 5530 Network Analyzer, http://www.alcatel.com/products/ productsummary.jhtml?relativePath=/com/en/appxml/opgproduct/ alcatel5530networkanalyzertcm228121701635.jhtml. [3] T. Starr, M. Sorbara, J. Cioffi, P. Silverman: DSL Advances, Upper Saddle River, Prentice Hall, 2002. [4] J. Verlinden, T. Bostoen, G. Ysebaert: Dynamic Spectrum Management for Digital Subscriber Lines, 2nd Edition, to be published 3rd Quarter 2005 on www.alcatel.com. [5] Self-managed DSL networks, Alcatel White Paper, to be published 3rd Quarter 2005 on www.alcatel.com.
Jan Verlinden is DSL Research Engineer in the Alcatel Access Networks Division, Antwerp, Belgium. (Jan.VJ.Verlinden@alcatel.be)
Reto Oehen is Product Line Manager in the Alcatel Access Networks Division, Antwerp, Belgium. (reto.oehen@alcatel.be)
www.alcatel.com/atr
159
TECHNICAL PAPER
OPERATIONAL EXCELLENCE
D. Rouffet, S. Kerboeuf, L. Cai, V. Capdevielle
4G MOBILE
4G will deliver low cost multi-megabit/s sessions any time, any place, using any terminal.
oice was the driver for second-generation mobile and has been a considerable success. Today, video and TV services are driving forward third generation (3G) deployment. And in the future, low cost, high speed data will drive forward the fourth generation (4G) as short-range communication emerges. Service and application ubiquity, with a high degree of personalization and synchronization between various user appliances, will be another driver. At the same time, it is probable that the radio access network will evolve from a centralized architecture to a distributed one.
Service Evolution
The evolution from 3G to 4G will be driven by services that offer better quality (e.g. video and sound) thanks to greater bandwidth, more sophistication in the association of a large quantity of information, and improved personalization. Convergence with other network (enterprise, fixed) services will come about through the high session data rate. It will require an always-on connection and a revenue model based on a fixed monthly fee. The impact on network capacity is expected to be significant. Machine-to-machine transmission will involve two basic equipment types: sensors (which measure parameters) and tags (which are generally read/write equipment). It is expected that users will require high data rates, similar to those on fixed networks, for data and stream160
ing applications. Figure 1: Service evolution vision Mobile terminal usage (laptops, Sender Personal digital Machine Human assistants, handReal helds) is expected time Video relay broadcasting VoIP to grow rapidly as Video supervision Videophone/conference Human navigation Interactive games they become more Internet browsing user friendly. Fluid Information service Chat Music download Visual mail/audio mail high quality video Text mail and network reacPermit delay tivity are important Location information Remote control user requirements. services, distribution Recording to storage devices: systems, etc. Key infrastructure Data transfer voice, video, etc. design requireConsumer electronic device maintenance ments include: fast response, high sesSupport for real-time and non-real-time services sion rate, high Source NTTDoCoMo capacity, low user charges, rapid return on investment for operators, investment that is in As a comparison, the expected best perline with the growth in demand, and sim- formance of 3G is around 10 bit/s/Hz/km2 ple autonomous terminals. using High Speed Downlink Packet Access The infrastructure will be much more (HSDPA), Multiple-Input Multiple-Output distributed than in current deployments, (MIMO), etc. No current technology is facilitating the introduction of a new capable of such performance. source of local traffic: machine-tomachine. Figure 1 shows one vision of Dimensioning objectives how services are likely to evolve; most Based on various traffic analyses, the such visions are similar. Wireless World Initiative (WWI) has issued target air interface performance figures. A consensus has been reached Dimensioning targets A simple calculation illustrates the Figure 2: Dimensioning examples order of magnitude. 10000 10000 Number of subscribers/km2 The design target in terms of radio Peak usage 20 % 20 % Spacial efficiency 30 % 20 % performance is to Various efficiencies (e.g. MAC) 40 % 30 % achieve a scalable Average service rate (kbit/s) 128 1000 capacity from 50 to Required capacity (Gbit/s/km2) 2.13 33.33 500 bit/s/Hz/km2 (including capacity Available bandwidth (MHz) 50 100 for indoor use), as 2 Required radio performances bit/s/Hz/km 42.7 333.3 shown in Figure 2.
Receiver Machine Human
www.alcatel.com/atr
4G MOBILE
around peak rates of 100 Mbit/s in mobile situations and 1 Gbit/s in nomadic and pedestrian situations, at least as targets. So far, in a 10 MHz spectrum, a carrier rate of 20 Mbit/s has been achieved when the user is moving at high speed, and 40 Mbit/s in nomadic use. These values will double when MIMO is introduced. Clearly, the bitrate should be associated with an amount of spectrum. For mobile use, a good target is a network performance of 5 bit/s/Hz, rising to 8 bit/s/Hz in nomadic use.
Figure 3: Multiple overlay architecture

Service adaptation Internet Gateway Router Slice manager PSTN Worldwide Internet Fixed networks radio Switch access controller access point Support Node(s) Satellite hub RNC NodeB
Macro mobility 4G Gateway
Enterprise networks access controller Gateway
Multi-technology Approach
Many technologies are competing on the road to 4G, as can be seen in Figure 3. Three paths are possible, even if they are more or less specialized. The first is the 3G-centric path, in which Code Division Multiple Access (CDMA) will be progressively pushed to the point at which terminal manufacturers will give up. When this point is reached, another technology will be needed to realize the required increases in capacity and data rates. The second path is the radio LAN one. Widespread deployment of WiFi is expected to start in 2005 for PCs, laptops and PDAs. In enterprises, voice may start to be carried by Voice over Wireless LAN (VoWLAN). However, it is not clear what the next successful technology will be. Reaching a consensus on a 200 Mbit/s (and more) technology will be a lengthy task, with too many proprietary solutions on offer. A third path is IEEE 802.16e and 802.20, which are simpler than 3G for the equivalent performance. A core network evolution towards a broadband Next Generation Network (NGN) will facilitate the introduction of new access network technologies through standard access gateways, based on ETSI-TISPAN, ITU-T, 3GPP, China Communication Standards Association (CCSA) and other standards. How can an operator provide a large number of users with high session data rates using its existing infrastructure? At least two technologies are needed. The first (called parent coverage) is dedicated to large coverage and real-time services. Legacy technologies, such as 2G/3G and their evolutions will be complemented by WiFi and WiMAX. A second set of technologies is needed to increase
Broadcast DVB-H DAB
access controller base station "hot spot/zone" slice WiMAX access point 4G coverage outdoor & indoor
Terrestrial repeater localization & broadcast slice
WiFi and indoor evolution
"cellular" slice UniRAN (3G+)
Terminal S/W
PAN AdHoc
DAB: Digital Audio Broadcasting DVB-H: Digital Video Broadcasting Handheld PAN: Personal Area Network PSTN: Public Switched Telephone Network RNC: Radio Network Controller
capacity, and can be designed without any constraints on coverage continuity. This is known as pico-cell coverage. Only the use of both technologies can achieve both targets (Figure 4). Handover between parent coverage and pico cell coverage is different from a classical roaming process, but similar to classical handover. Parent coverage can also be used as a back-up when service delivery in the pico cell becomes too difficult.
Key 4G Technologies
Some of the key technologies required for 4G are briefly described below:
Figure 4: Coverage performance trends
>8U
OFDMA Orthogonal Frequency Division Multiplexing (OFDM) not only provides clear advantages for physical layer performance, but also a framework for improving layer 2 performance by proposing an additional degree of free-
Pico cell A good example of a pico cell is a WiFi coverage. By extension, a pico cell has a radius around 50 m and the associated base station is similar to a WiFi access point. It can be deployed indoors or outdoors.
10-25 U 15 to 40 U > 500 U
Bearer oriented
3G
3G+
3G evolution
25 U Packet and Bearer oriented
50 U WiMAX evolution
Parent coverage and pico cell coverage
WiMAX
4G
Nano coverage
100 U Packet oriented WiFi SRR
> 500 U WiFi SRR evolution U = bit/s/Hz/km2
> 1000 U
New bands
New sites
www.alcatel.com/atr
161
4G MOBILE
Figure 5: OFDM principles
10 MHz Bandwidth FFT Guard Intervals Symbols Frequency Sub-carriers
eral modems at a given Base Transceiver Station (BTS). SDR makes this reconfiguration easy. In the context of 4G systems, SDR will become an enabler for the aggregation of multi-standard pico/micro cells. For a manufacturer, this can be a powerful aid to providing multi-standard, multi-band equipment with reduced development effort and costs through simultaneous multi-channel processing. Multiple-input multiple-output MIMO uses signal multiplexing between multiple transmitting antennas (space multiplex) and time or frequency. It is well suited to OFDM, as it is possible to process independent time symbols as soon as the OFDM waveform is correctly designed for the channel. This aspect of OFDM greatly simplifies processing. The signal transmitted by m antennas is received by n antennas. Processing of the received signals may deliver several performance improvements: range, quality of received signal and spectrum efficiency. In principle, MIMO is more efficient when many multiple path signals are received. The performance in cellular deployments is still subject to research and simulations (see Figure 6). However, it is generally admitted that the gain in spectrum efficiency is directly related to the minimum number of antennas in the link. Interlayer optimization The most obvious interaction is the one between MIMO and the MAC layer. Other interactions have been identified (see Figure 7). Handover and mobility Handover technologies based on mobile IP technology have been considered for data and voice. Mobile IP techniques are slow but can be accelerated with classical methods (hierarchical, fast mobile IP). These methods are applicable to data and probably also voice. In single-frequency networks, it is necessary to reconsider the handover methods. Several tech-
Time
FFT: Fast Fourier transform
Figure 6: Alcatel test-bed performance roadmap

Mbps 20 MHz 8 bit/s/Hz 200 10 MHz Larger bandwidth MAC & MIMO joint optimization Robust MIMO to correlation and Doppler 5 bit/s/Hz 100 MIMO-OF DM MAC layer OFDM 50 20 2 bit/s/Hz 20 Mbit/s 2004 MAC: Media Access Control 4 bit/s/Hz 40 Mbit/s 20 Mbit/s 2005 40 Mbit/s Test bed development 100 Mbit/s 100 Mbit/s 160 Mbit/s Air interface design and verification by simulation MIMO-OFDM optimization: higher order modulation, improved channel coding Advanced MIMO detector
2006
2007
dom. Using ODFM, it is possible to exploit the time domain, the space domain, the frequency domain and even the code domain to optimize radio channel usage. It ensures very robust transmission in multi-path environments with reduced receiver complexity. As shown in Figure 5, the signal is split into orthogonal subcarriers, on each of which the signal is narrowband (a few kHz) and therefore immune to multi-path effects, provided a guard interval is inserted between each OFDM symbol. OFDM also provides a frequency diversity gain, improving the physical layer performance. It is also compatible with other enhancement technologies, such as smart antennas and MIMO. OFDM modulation can also be employed as a multiple access technology (Orthogonal Frequency Division Multiple Access; OFDMA). In this case, each OFDM symbol can transmit information to/from several users using a different set of subcarriers (subchannels). This not only provides additional flexibility for
162
resource allocation (increasing the capacity), but also enables cross-layer optimization of radio link usage.
Software defined radio Software Defined Radio (SDR) benefits from todays high processing power to develop multi-band, multi-standard base stations and terminals. Although in future the terminals will adapt the air interface to the available radio Figure 7: Layer interaction and associated optimization access technology, at present this is OFDMA done by the infra Smart antennas structure. Several MIMO infrastructure gains Optimized MAC scheduling algorithms (e.g. MUD) are expected from Robust and scalable transmissions SDR. For example, Network selection (e.g. ABC) to increase network Usage of cache capacity at a spe Pico station and associated deployment cific time (e.g. during a sports event), IP mobility an operator will Meshed networks and ad hoc routing reconfigure its network adding sev- ABC: Always Best Connected MUD: Multi User Detection
www.alcatel.com/atr
4G MOBILE
niques can be used when the carrier to is shown in Figure 8. At the entrance of the interference ratio is negative (e.g. VSF- access network, lines of cache at the OFDM, bit repetition), but the drawback destination of a terminal are built and of these techniques is capacity. In OFDM, stored. When a terminal enters an area in the same alternative exists as in CDMA, which a transfer is possible, it simply asks which is to use macro-diversity. In the case for the line of cache following the last of OFDM, MIMO allows macro-diversity received. Between the terminal and the processing with performance gains. However, the implementation of macro-diversity Figure 8: Pico cell network design implies that MIMO processing is centralized and transmissions are synchronous. This is not as complex as in CDMA, IP Backbone but such a technique should Cache only be used in situations making IP where spectrum is very Cache scarce. Proxy
G/W L1/L2
Caching and Pico Cells

Memory in the network and terminals facilitates service delivery. In cellular systems, this extends the capabilities of the MAC scheduler, as it facilitates the delivery of real-time services. Resources can be assigned to data only when the radio conditions are favorable. This method can double the capacity of a classical cellular system. In pico cellular coverage, high data rate (non-real-time) services can be delivered even when reception/transmission is interrupted for a few seconds. Consequently, the coverage zone within which data can be MT: Mobile received/transmitted can be designed with no constraints other than limiting interference. Data delivery is preferred in places where the bitrate is a maximum. Between these areas, the coverage is not used most of the time, creating an apparent discontinuity. In these areas, content is sent to the terminal cache at the high data rate and read at the service rate. Coverages are discontinuous. The advantage of coverage, especially when designed with caching technology, is high spectrum efficiency, high scalability (from 50 to 500 bit/s/Hz), high capacity and lower cost. A specific architecture is needed to introduce cache memory in the network. An example
Uni or N-Casting
4G d-cov AN
Cache L1/L2
Cache storage
ently: one base layer (30 kbit/s), which is a robust flow but of limited quality (e.g. 5 images/s), and two enhancement flows (50 kbit/s and 200 kbit/s). The first flow provides availability, the other two quality and definition. In a streaming situation, the terminal will have three caches. In pico cellular coverage, the parent coverage establishes the service dialog and service start-up (with the base layer). As soon as the terminal enters pico cell coverage, the terminal caches are filled, starting with the base cache. Video (and audio) transmissions are currently transmitted without error and without packet loss. However, it is possible to allow error rates of about 10-5 /10-6 and a packet loss around 10-2 /10-3. Coded images still contain enough redundancy for error correction. It is possible to gain about 10 dB in transmission with a reasonable increase in complexity. Using the described technologies, multimedia transmission can provide a good quality user experience. Coverage Coverage is achieved by adding new technologies (possibly in overlay mode) and progressively enhancing density. Take a WiMAX deployment, for example: first the parent coverage is deployed; it is then made denser by adding discontinuous pico cells, after which the pico cell is made denser but still discontinuously. Finally the pico cell coverage is made continuous either by using MIMO or by deploying another pico cell coverage in a different frequency band (see Figure 9). The ultimate performances of the various technologies are shown in Figure 10. Parent coverage performance may vary from 1 to 20 bit/s/Hz/km?, while pico cell technology can achieve from 100 to 500
RAP L1 MAC Cache MT IP
Pico cell network design Terminal
cache. A simple, robust and reliable protocol is used between the terminal and the cache for every service delivered in this type of coverage. Multimedia service delivery, service adaptation and robust transmission Audio and video coding are scalable. For instance, a video flow can be split into three flows which can be transported independ-
Figure 9: Example of deployment in dense traffic areas
Parent coverage
Parent coverage with discontinuous pico cells
Continuous pico cells coverage with MIMO
www.alcatel.com/atr
163
4G MOBILE
Distributed architecture An architecture with a large number of decentralized connections to the core network. Network selection The process of associating one or more service networks with a user terminal.
ETSI- TISPAN). This generic architecture integrates all service enablers (e.g. IMS, network selection, middleware for applications providers), and offers a unique interface to application service providers.
Conclusion
The provision of megabit/s data rates to thousands of radio and mobile terminals per square kilometer presents several challenges. Some key technologies permit the progressive introduction of such networks without jeopardizing existing investment. Disruptive technologies are needed to achieve high capacity at low cost, but it can still be done in a progressive manner. The key enablers are: Sufficient spectrum, with associated sharing mechanisms. Coverage with two technologies: parent (2G, 3G, WiMAX) for real-time delivery, and discontinuous pico cell for high data rate delivery. Caching technology in the network and terminals. OFDM and MIMO. IP mobility. Multi-technology distributed architecture. Fixed-mobile convergence (for indoor service). Network selection mechanisms. Many other features, such as robust transmission and cross-layer optimization, will contribute to optimizing the performance, which can reach between 100 and 500 bit/s/Hz/km 2. The distributed, full IP architecture can be
Denis Rouffet is Head of Strategic Projects in the Research & Innovation department, Chief Technology Office, Marcoussis, France. (Denis.Rouffet@alcatel.fr)
bit/s/Hz/km?, depending on the complexity of the terminal hardware and software. These performances only refer to outdoor coverage; not all the issues associated with indoor coverage have yet been resolved. However, indoor coverage can be obtained by: Direct penetration; this is only possible in low frequency bands (significantly below 1 GHz) and requires an excess of power, which may raise significant interference issues. Indoor short range radio connected to the fixed network. Connection via a relay to a pico cellular access point. Integration in a Broadband NGN The focus is now on deploying an architecture realizing convergence between the fixed and mobile networks (ITU-T Broadband NGN and
Sylvaine Kerboeuf is leading a research group in the Research & Innovation department, Chief Technology Office, Marcoussis, France. (Sylvaine.Kerboeuf@alcatel.fr)
Liyu Cai is leading a research group in the Research & Innovation department, Chief Technology Office, Alcatel Shanghai Bell, Shanghai, China . (Liyu.Cai@alcatel-sbell.com.cn)
Vronique Capdevielle is leading a research group in the Research & Innovation department, Chief Technology Office, Marcoussis, France . (Veronique.Capdevielle@alcatel.fr) deployed using two main products: base stations and the associated controllers. Terminal complexity depends on the number of technologies they can work with. The minimum number of technologies is two: one for the radio coverage and one for short range use (e.g. PANs). However, the presence of legacy networks will increase this to six or seven.
164
www.alcatel.com/atr
CUSTOMER APPLICATION NOTE
CUSTOMER EXAMPLES
O. Perot, J-M. Perera, L. Byerley
3G NETWORK POWERS INTENSE SERVICES

The launch of Orange Intense marks the culmination of three successful years of close collaboration with Alcatel.
ased on Alcatels Evolium Third Generation Universal Mobile Telecommunications System (3G/UMTS) radio, core network and multimedia solutions, on 6th December 2004, Orange France launched Orange Intense, its first 3G mobile telephony service. These solutions are also the basis of a major 3G launch by Orange in the UK, to be followed by a similar launch in the Netherlands in 2005. The Orange France 3G launch marks only the second major mobile communication technology to be rolled out in an entirely new network since the advent of mobile telephony in France. Indeed, this completely new network represents a quantum leap forward from the narrowband Global System for Mobile communication (GSM) to the broadband UMTS from voice to rich user-centric multimedia services.
Quantum Leap
Although this quantum leap didnt take place in the blink of an eye, for a project of this size the launch has been achieved at a remarkable pace by any time-to-market standard. When one considers that the Orange network now offers the latest and most exciting UMTS-based multimedia services to approximately 40% of the French population, the scale of the project can be quickly grasped. While participants and observers have attributed this success to a diverse range of factors, including careful planning, laboratory experiments, field trials, large-scale pilot projects, and extensive market and user testing, they all agree that after three years of close collaboration, Alcatel and Orange have raised the bar to an unprecedented level for the communication industry worldwide.
Contributing to the challenge was the complexity of the mobile communications ecosystem. As an operator / service provider, Orange was surrounded by numerous participants with many offerings and goals. These participants included editors, audiovisual content producers, content aggregators, audio-visual rights holders, games producers, and a community of developers. Prioritizing these partnerships, negotiating revenue sharing, and guaranteeing an interoperable platform for third-party services were just a few of the challenges. No surprise, then, that Orange France chose first time right as the guiding motto for its 3G network and multimedia services launch. Applying this motto, Orange embarked upon a deliberate course of action, starting with the careful selection of suppliers and partners. Then, working closely with its trusted partners, Orange focused efforts on two fronts: network infrastructure and technologies, and service definition and market development. Alcatel played a crucial role in both these areas during the 3G services launch.
Building the Partnership

By the end of 2001, Alcatel had become much more than a mere supplier; it had become a true partner of Orange France. The two marketing teams shared many of the same values and goals, and worked together to meet the new challenges of 3G. For example, the early engagement, which was marketing focused, enabled Alcatel to analyze and design a new set of 3G services. This, in turn, enabled it to identify appropriate partners, including content, application and terminal providers, some of whom Alcatel managed directly on behalf of Orange. Among other tasks and services, Alcatel managed the integration and hosting of most of the multimedia services and the Orange World Video Portal. This was done with Atomiz, a French company dedicated to enhancing the quality of video streams for mobile users, and with content providers such as Dioranews, MCM for music, and BeTomorrow for online games. Equally importantly, Alcatel demonstrated the viability of 3G services in comprehensive trials at its 3G Reality Centers, which have been designed by Alcatel to develop and test mobile solutions in a live end-to-end environment. The trialing expertise proved crucial for the early development and validation of new and innovative 3G services. The 3G Reality Centers also provided comprehensive expertise in interoperability between terminals, applications and network elements.
165
Shared Vision
As one of the worlds largest mobile communication companies, and a subsidiary of the France Telecom group, the Orange Group realized that it needed a partner that shares its vision of a true 3G broadband world one in which users could enjoy the benefits of a wide variety of life-enhancing and personalized services. The stakes, moreover, were high. With operations in 17 countries across Europe and beyond (including 13.9 million active customers in the UK and more than 20 million registered customers in France), Orange has 52 million customers worldwide. Deploying a UMTS network required that Orange find a partner with the experience and innovation, solutions integration, and technical expertise to make its vision come alive.
www.alcatel.com/atr
Not surprisingly, Alcatels service-oriented approach began to deliver even more fruit in mid-2002, when Orange France offered its first end-to-end RFQ for a 3G pilot project called Ville Orange. Later that year, Alcatel was contracted to provide all 3G network and live services, as well as the integration and hosting for the city trials. This project was successfully managed and delivered in 2003. Orange and Alcatel share the same objectives - delivering services to customers first. Thats what we have achieved with 3G right now. Bruno Dachary, Mobile Broadband Consumer Launch Director, Orange France Table 1 summarizes the major milestones for the project.
Table 1: Major milestones

Alcatel wins the first 3G Framework Agreement with the Orange Group to supply UMTS Terrestrial Radio Access Network (UTRAN) solutions for the Orange Group 3G deployment in Europe. The Alcatel UTRAN is installed in France, UK and The Netherlands. Together, Alcatel and Orange France run the first live 3G calls in Paris over an Evolium UTRAN and the Alcatel Evolium core network. Alcatel begins working with the Orange consumer marketing team to define the 3G multimedia services and how to launch them first time right! The Ville Orange 3G pilot project is initiated by Orange. Alcatel is contracted to provide all 3G network and live services, integration and hosting for city trials. The Ville Orange trial is conducted in 12 cities for up to 7000 subscribers. Alcatel wins the Second Orange Group Framework Agreement for UTRAN. It is selected in France to deliver half the network, including Paris and the environs, as well as to cover one-fifth of the coverage area in the UK. Two-thousand 3G/UMTS radio sites in France are rolled out and commissioned. As a result of the Ville Orange project, Alcatel wins and delivers three multimedia projects awarded by Orange for the commercial launch, including a video streaming solution, a fixed mobile video gateway, and a video mail box. Final acceptance of the commercial launch of the UTRAN release, and end-to-end optimization of the 3G network from terminals to applications is achieved. The commercial launch of Orange Intense begins.
Mid 2001
October 2001 December 2001 July 2002 December 2002 April to October 2004
July 2003 September 2003 to May 2004
April to October 2004
Road from Ville Orange to the Commercial Launch
June to November 2004
In 2004, the Ville Orange 3G pilot December 2004 test was conducted in twelve French towns: Toulouse, Lille, Cannes, Nice, Lens, Nantes, Bordeaux, Lyon, Marseille, Strasbourg, Rennes and Poitiers (see Figure 1). The trial involved a large-scale, user-centric test not only for the network and related services, but also for distribution, customer service, billing, service ergonomics and usage, as well as after-sales service. Between five and seven thousand pub-
Figure 1: Cities included in the Ville Orange 3G Pilot Project

Lille Lens
Strasbourg Rennes
FRANCE
Nantes Poitiers
Bordeaux Toulouse Nice Cannes Marseille
lic and business customers were able to test the 3G services, and to respond to the finely-tuned Orange Intense offer which was the basis for the commercial launch. Services tested for Ville Orange included videotelephony, fixed-to-mobile video telephony, web conferencing, video mailbox, online games, web cam access, video streaming, live TV, live sports events, personal multimedia storage, instant messaging, video Multimedia Messaging Service (MMS) and content management related to the Ville Orange services. As part of the test project, Alcatel supplied the 3G network infrastructure integration, optimization and ongoing operations and maintenance support. Alcatel was also responsible for ensuring the interoperability and end-to-end integration of the UMTS Terrestrial Radio Access Network (UTRAN) with selected Orange terminals, as well as with the existing circuit and packet core networks. On the software and services side, Alcatel oversaw the endto-end network and service integration, guaranteeing Quality of Service (QoS) as well as performance from terminal to content. Just as important, the complete solution required that Alcatel integrate every member of the Orange ecosystem into the Orange information systems from application enablers to content providers. For some services, Alcatel also provided turnkey hosting, as well as operating the services and related platforms.
Partner for Orange Intense

Alcatels collaboration on Ville Orange proved so successful that it was asked to be a partner on Orange Intense the final phase or commercial launch. To meet the challenge of first time right, Alcatel conducted a rigorous and demanding acceptance program on the radio access and core net-
166
www.alcatel.com/atr
works to ensure they provided optimum end-to-end performance from terminals to transmission. True to form, Alcatels operational and technical teams stayed on schedule and achieved the QoS benchmarks in a complex mobile environment. In fact, Alcatel installed 800 radio base stations (Node Bs) in just three months, enabling Orange to meet its ambitious time-to-market goal, launching Orange Intense on 6th December 2004. Broadband Services Tuned to Market Demand Orange Intense offers subscribers a complete portfolio of new mobile multimedia services, such as audio and video streaming, videomail box, and convergent fixed/mobile video calls. The Orange World Video Portal is at the heart of this offering. With its easy-to-use menu, the portal delivers new multimedia audio, video and television Figure 2: The Orange World video portal services, making it unique in the mobile market (see Figure 2). To ensure a finely tuned service, an upgraded portal has been delivered to Orange every two months since 2003. This new video portal is completely different from what exists in the market, says Bruno Dachary, Orange France. Its very simple to use and is based only on pictures. To access a video, you just click on a picture once or twice, so its very straightforward for customers. With Alcatel, we are really focusing on what the customer needs. On this project, Alcatel served as the video portals endto-end integrator, as well as the exclusive supplier of all streaming technology. The streaming technology, including the server, has been developed by PacketVideo Network SoluFigure 3: Alcatel-developed video tions, a wholly owned mailbox interface subsidiary of Alcatel. With Orange Intense, subscribers can access more than 20 television channels, live web cams and many audio and video streams. These include news, comedy, sports, fashion, people, movies, cartoons and music. Among the services, the video mailbox is proving especially attractive with about twenty percent of customers opting for the service. With video mailbox (see Figure 3),
when a called party is absent, callers can record a video message using either a camera-enabled mobile phone or a desktop web cam stored in the users video mailbox. Messages can also be sent from one video mailbox to another. Figure 4 shows a typical video call.
Why Alcatel?
Orange France chose Alcatel as its strategic partner for the delivery of its 3G network based on Alcatels ability to repeatedly meet Oranges deadlines. Moreover, Alcatel successfully completed a complex end-to-end acceptance process, resulting in optimized, stable and proven software. Throughout the project, the trust placed in Alcatel proved well founded. Whether as the preferred integrator of the Orange World Video Portal, the provider of turnkey services, or the validator of interoperability between all the elements of the end-to-end solution, Alcatel has delivered on the Orange vision every step of the way. As Dachary observes: The advantage of working with a partner like Alcatel is that we are quite close. We have been working together for a long time, so we know each other quite well. This creates some good links and good ways to work towards the same target. Other factors were also crucial in Oranges choice of Alcatel. Technologically, the Evolium UTRAN product line, developed through the Alcatel-Fujitsu joint Figure 4: Typical Orange Intense video call venture, was the perfect combination of Fujitsus stable and proven platform and Alcatels expertise in building the most efficient and cost-effective radio base stations. Alcatels pioneering vision of open and interoperable solutions, originally developed for the General Packet Radio Service (GPRS), enabled Alcatel to meet Oranges requirement for guaranteed, end-toend quality of service. Finally, Alcatels operational capability had proved itself in the past during a long-standing GSM relationship with Orange. This capability included: mobilization for ontime delivery, reactivity to feedback and fast correction, optimization know-how, and worldwide delivery and technical support. The trailblazing continues Since the launch of Orange Intense in France, Alcatel has become the first French company equipped with 3G indoor technology. Indoor coverage is becoming essential for new multimedia services being introduced with 3G. The services require extremely powerful equipment and excellent QoS. Not surprisingly, when Alcatel decided to equip its Paris headquarters with 3G indoor coverage, it chose Orange France as its 3G network. Orange increased its 3G radio coverage at rue La Botie via an indoor 3G distribution network
167
www.alcatel.com/atr
supplied by Alcatel. Now from the basement to the top floor, and even in the elevators, Alcatel employees can make 3G calls without any fear that they will be dropped. They can also make video calls and access the latest multimedia services. Once again, Alcatels expertise in 3G has been the key success factor in this premier 3G indoor installation.
Conclusion
From conception to execution of the Orange 3G network and multimedia services, Alcatel and Orange have proved successful partners. With both companies committed to a similar vision of new and innovative 3G services, they have been able to leverage each others strengths every step of the way. The net result is an unprecedented 3G network, rolled out at an unprecedented speed.
Olivier Perot was Vice President of the Orange Support Center in the Alcatel Mobile Communications Group, Vlizy, France, until 1st March 2005. He is now in charge of the Regional Support Center covering Area 4 within the Fixed Solutions Division. (Olivier.perot@alcatel.fr)
Jean-Marc Perera is Director of Marketing & Solutions in the Orange Support Center of the Alcatel Mobile Communications Group, Vlizy, France. (Jeanmarc.perera@alcatel.fr)
Lorraine Byerley is Head of Web Communications for the Alcatel Mobile Communications Group in Ottawa, Canada. (lorraine.byerley@alcatel.com)
168
www.alcatel.com/atr
EDITORIAL OFFICES
ENGLISH EDITION Catherine Camus

Alcatel Telecommunications Review Alcatel
Route de Nozay 91461 Marcoussis Cedex - France Tel.: 33 (0)1 69 63 16 49 - Fax: 33 (0)1 69 63 42 50 E-mail: catherine.camus@alcatel.fr
Publisher: Jean-Paul BARTH - Technical Journal published by Compagnie Financire Alcatel, a corporation (Socit Anonyme) with a share capital of Eur 2 610 910 922. Registered office: 54, rue de la Botie, 75008 Paris - France Trade registration: RCS Paris B 351 213 624 Major shareholder: Alcatel (99.9%) Legal registration: June 2005 ISSN: 1267-7167.
Illustrations and layout by Atelier Antoine Maiffret (www.maiffret.net) Cover photography: Corbis
Associate Editors
Printed in France by: MACON Imprimerie 22, rue du 134e Rgiment dInfanterie - 71000 MACON. English Edition: 26 000 copies.
SPANISH EDITION Jose-Luis Femenia Gonzalez

Revista de Telecomunicaciones de Alcatel Alcatel
Ramirez de Prado 5 - 28045 Madrid - Spain Tel.: (34-91) 330 46 14 E-mail: Jose_Luis.Femenia_Gonzalez@alcatel.es
Compagnie Financire Alcatel
CHINESE EDITION Yuan Hong

Communications Department Alcatel Shangai Bell Co, Ltd
388#, Ningqiao Road, Pudong Jinqiao Shanghai 201206 - Peoples Republic of China Tel.: +8621 58541240 Fax: +8621 58540791/58545951 E-mail: Hong.h.Yuan@alcatel-sbell.com.cn
UNIVERSAL ACCESS
Editorial: Ernie Carey Introduction: M. Peruyero, Y. TJoens UNIVERSAL BROADBAND ACCESS I Making User-Centric Broadband in Access a Reality ANY ACCESS I Universal Broadband Access: Going Wireless and Mobile I Software Defined Radio: A Promising Technology for Multi-Standard Base Stations I Standardization: Key to Mass Deployment I Optical Fibers Pave the Way to Faster Broadband Access I Extending Broadband Reach by Satellite ANY SERVICE I Supporting Quality of Service in Broadband Access Networks I Access Network Enhancements for the Delivery of Video Services CONSOLIDATION I Fixed Access Vision I WiMAX: From Fixed Wireless Access to Internet in the Pocket I Alcatel Evolium Multi-standard Radio Access System OPERATIONAL EXCELLENCE I Optimizing DSL for Multimedia Services I 4G Mobile CUSTOMER EXAMPLES I 3G Network Powers Intense Services
Next Issue:
DATA-AWARE TRANSPORT
Paper and cover chlorine-free - Cover: Corbis

ALCATEL ConnectMe Broadband

Uploaded by

Document Information

Original Description:

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

ALCATEL ConnectMe Broadband

Uploaded by

Copyright:

Available Formats

CONNECT ME ANYWHERE, ANY TIME, ANY WAY ...

ALCATEL TELECOMMUNICATIONS REVIEW

Alcatel Telecommunications Review - 2 nd Quarter 2005

Connect me anywhere, any time, any way

Universal Broadband Access

Ph. Lain, L. Le Gouriellec, J. De Vriendt

B. Haberland, W. Koenig, A. Pascht, U. Weiss Standardization: Key to Mass Deployment

Th. Pfeiffer, E. Ringoot, A. Granger, D. Wang

Articles marked @ are the web supplement to this edition at www.alcatel.com/atr

Extending Broadband Reach by Satellite

Alcatel Evolium Multi-standard Radio Access System

Operational Excellence Any Service

T. Bostoen, R. Oehen, J. Verlinden 4G Mobile

D. Rouffet, S. Kerboeuf, L. Cai, V. Capdevielle

R. Sharpe, D. Zriny, D. De Vleeschauwer

O. Perot, J-M. Perera, L. Byerley

D. Renaudeau, D. Boettle, H. Steyaert

Customer Applications Notes www.alcatel.com/atr

Technology White Paper

Strategy White Paper

2 nd Quarter 2005 - Alcatel Telecommunications Review -

SBC PROJECT LIGHTSPEED

- Alcatel Telecommunications Review - 2 nd Quarter 2005

SBC PROJECT LIGHTSPEED

2 nd Quarter 2005 - Alcatel Telecommunications Review -

Michel Peruyero, Yves TJoens

Figure 1: Universal broadband access

Every Terminal, Everywhere

Universal Broadband Access Wireline

Wireless Satellite Mobile

Service Aware Edge & Data Aware Transport

Open Services Delivery Environment

Operations & Business Support Systems Integration & Operations

- Alcatel Telecommunications Review - 2 nd Quarter 2005

Michel Peruyero, Yves TJoens

2 nd Quarter 2005 - Alcatel Telecommunications Review -

STRATEGY WHITE PAPER

UNIVERSAL BROADBAND ACCESS

Ph. Lain, L. Le Gouriellec, J. De Vriendt

MAKING USER-CENTRIC BROADBAND IN ACCESS A REALITY

- Alcatel Telecommunications Review - 2 nd Quarter 2005

MAKING USER-CENTRIC BROADBAND IN ACCESS A REALITY

Figure 2: Five services classes and their requirements

One or More Access Networks

Fixed: DSL, Cable, etc.

Access Network (3G, WiFi/WiMAX, Fixed DSL, Cable) Application

Requirements To Access Networks

2 nd Quarter 2005 - Alcatel Telecommunications Review -

MAKING USER-CENTRIC BROADBAND IN ACCESS A REALITY

Offer from the Operators

WiFi/DSL, cable, Sat at home

WiFi hot spots WiMAX hot zone

- Alcatel Telecommunications Review - 2 nd Quarter 2005

MAKING USER-CENTRIC BROADBAND IN ACCESS A REALITY

GPRS: General Packet Radio Service

2 nd Quarter 2005 - Alcatel Telecommunications Review -

MAKING USER-CENTRIC BROADBAND IN ACCESS A REALITY

- Alcatel Telecommunications Review - 2 nd Quarter 2005

STRATEGY WHITE PAPER

J-L.Hurel, J. Brouet, L. Le Gouriellec, M. Peruyero

UNIVERSAL BROADBAND ACCESS: GOING WIRELESS AND MOBILE

key access technologies

Driving Forces behind Broadband Wireless and Mobile