You are on page 1of 35

IBM Systems and Technology Group

SAN Volume Controller Release 4.2 and 4.3 Performance


v

Bruce McNutt

2008 IBM Corporation

IBM Systems and Technology Group

SVC Facts
IBM has 40 years experience in virtualization technologies IBM has shipped over 13,000 SVC engines running in more than 4,300 SVC systems There are more than 130 customer references and 24 customer case studies for SAN Volume Controller SAN Volume Controller is a proven offering that has been delivering benefits to customers for four years SAN Volume Controller demonstrates scalability with the fastest Storage Performance Council benchmark results SAN Volume Controller can virtualize IBM and non-IBM storage (over 120 systems from IBM, EMC, HP, HDS, Sun, Dell, NetApp, Fujitsu, NEC, Bull)

2008 IBM Corporation

IBM Systems and Technology Group

Whats New with SAN Volume Controller Version 4


Space-Efficient VDisks / thin provisioning (4.3) Incremental FlashCopy; FlashCopy with multiple and/or cascaded targets (4.2.1) 4Gbps fabric support (4.1)
Improves infrastructure simplification through use of latest SAN technologies Extends business continuity reach to practically unlimited distances

Global Mirror (4.1.1) Dual core processing platform (4.2) Improved performance statistics (4.1)
Enables better monitoring, understanding, and planning of SVC systems Foundation for continued growth and technology exploitation Extends range of environments supported by SVC

Cluster non-disruptive upgrade capability (4.1) Additional server and disk system support (ongoing)

2008 IBM Corporation

IBM Systems and Technology Group

And by the way, performance


Performance enables consolidation. High performance enables the virtualization of big environments. SVC has it in spades.
The top performing disk storage product from any vendor, by a wide margin.

2008 IBM Corporation

IBM Systems and Technology Group

Storage Performance Council Benchmarks


Storage Performance Council (SPC) www.storageperformance.org

SPC-1
Random Workload
Dispersed small data Database queries OLTP

SPC-2
Sequential Workload
Consecutive large data Database table scanning Video, streaming media

Relevant Industries
Retail Financial Governmental

Relevant Industries
Media/Entertainment Data Mining Life/Physical Sciences

2008 IBM Corporation

IBM Systems and Technology Group

SPC-1 results through 2/1/2008 (most recent first)


0
T e xa s M e m o ry S ys t e m s R a m S a n- 4 0 0 E M C C la riio n C X - 3 M o de l 4 0 ( wit h s na pv ie w) E M C C la riio n C X - 3 M o de l 4 0 ( no s na pv ie w) N e t A pp F A S 3 0 4 0 ( wit h s na ps ho t ) N e t A pp F A S 3 0 4 0 ( no s na ps ho t ) H D S US P - V , S un 9 9 9 0 - V , H P X P 2 4 0 0 0 F S S - E T E R N US 8 0 0 0 M o de l 110 0 IB M S V C 4 .2 S un S t o ra ge T e k 6 5 4 0 IB M D S 4 8 0 0 IB M D S 8 3 0 0 T urbo S un S t o ra ge T e k 6 14 0 F S S - E T E R N US 4 0 0 0 M o de l 5 0 0 IB M D S 4 7 0 0 E xpre s s M o de l 7 2 F S S - E T E R N US 6 0 0 0 M o de l 110 0 IB M T o t a lS t o ra ge D S 8 3 0 0 IB M T o t a lS t o ra ge S A N V o lum e C o nt ro lle r 3 .1 IB M T o t a lS t o ra ge D S 4 8 0 0 S t o ra ge T e k F le xLine T M F LX 3 8 0 S t o ra ge S ys t e m S un S t o rE dge T M 6 13 0 A rra y S un S t o rE dge T M 6 9 2 0 S un S t o rE dge T M 3 5 10 IB M T o t a lS t o ra ge S A N V o lum e C o nt ro lle r 1.2 .1 D a t a C o re S A N m e lo dy D is k S e rv e r F S S - E T E R N US 6 0 0 0 M o de l 110 0 S un S t o rE dge T M 6 9 2 0 ( 8 t ra y) S un S t o rE dge T M 6 9 2 0 ( 2 0 t ra y) F S S - E T E R N US 3 0 0 0 M o de l 7 0 0 F S S - E T E R N US 3 0 0 0 M o de l 10 0 IB M T o t a lS t o ra ge S A N V o lum e C o nt ro lle r 1.1.1 D a t a C o re S A N m e lo dy D is k S e rv e r T e xa s M e m o ry S ys t e m s R a m S a n- 3 2 0 3 P A R InS e rv e S 8 0 0 X - S e rie s F S S - E T E R N US 6 0 0 0 M o de l 8 0 0 S t o ra ge T e k D 2 8 0 D is k S ys t e m ( m irro re d) S t o ra ge T e k D 2 8 0 D is k S ys t e m ( no n- m irro re d) S t o ra ge T e k D 2 4 0 D is k S ys t e m ( m irro re d) S t o ra ge T e k D 2 4 0 D is k S ys t e m ( no n- m irro re d) S un S t o rE dge T M 3 5 10 ( Wit hdra wn) IB M T o t a lS t o ra ge F A S t T 9 0 0 ( no n- m irro re d) IB M T o t a lS t o ra ge F A S t T 9 0 0 ( m irro re d) IB M T o t a lS t o ra ge F A S t T 6 0 0 w/ T urbo ( nm ) IB M T o t a lS t o ra ge F A S t T 6 0 0 w/ T urbo ( m ) F S S - E T E R N US 3 0 0 0 M o de l 6 0 0 M D a t a C o re S A N s ym pho ny N e t wo rk E dit io n F S S - E T E R N US 3 0 0 0 M o de l 6 0 0 S un S t o rE dge 6 3 2 0 ( 2 2 t ra y) S un S t o rE dge 6 3 2 0 ( 12 t ra y) F S S - E T E R N US 3 0 0 0 M o de l 4 0 0 3 P A R InS e rv e S 8 0 0 ( wit hdra wn) IB M T o t a lS t o ra ge E nt e rpris e S t o ra ge S e rv e r 8 0 0 H P - E nt e rpris e V irt ua l A rra y 2 C 12 D ( m irro re d) 3 P A R InS e rv S 8 0 0 ( wit hdra wn) H P - E nt e rpris e V irt ua l A rra y 2 C 12 D ( no n- m irro re d) D e ll- P E R C 3 / Q C S C S I R A ID C o nt ro lle r S un S t o rE dge 9 9 10 LS I E 4 6 0 0 F ibre C ha nne l S t o ra ge S ys t e m IB M T o t a lS t o ra ge E nt e rpris e S t o ra ge S e rv e r F 2 0

50,000

100,000

150,000

200,000

250,000

300,000

SVC is the fastest disk storage system

IBM disk storage Other disk storage Electronic Drum

2008 IBM Corporation

IBM Systems and Technology Group

Configuration for SVC 4.2 and 4.3 Benchmarks


SVC Node Pair

Switch 1

Switch 2

Switch 3

Switch 4

DS 4700

DS 4700

DS 4700

DS 4700

to host (s)
Represents 4 FC paths per line drawn Represents 8 FC paths per line drawn

Each DS4700 attaches five expansion drawers, for a total of 1536 15K RPM disks.
7 2008 IBM Corporation

IBM Systems and Technology Group

Miscellaneous Benchmark Settings


Component SVC SVC SVC SVC SVC DS4700 DS4700 DS4700 DS4700 DS4700 AIX AIX
8

Setting Extent Size (MB) Management Mode Space-Efficient Grain Size Flash Copy Grain Size Flash Copy Clean Rate Segment Size (KB) Cache Block Size (KB) Readahead Cache Mirroring RAID-10 Array Size hdisk queue depth LVM policy

Value 256 Sequential 256 64 0 256 16 1 Enabled 12 or 10+S+S 40 striped


2008 IBM Corporation

IBM Systems and Technology Group

Performance of the virtualization layer


OLTP 100 80 WM 80 60 40 20 50 100 150 WH 50 100 200 300 100 400 500 RH 150 600 200 250 RM Base 4300 Base 4300+SVC 3.1 40 20 50 100 60 70/30/50 150

OLTP 100 80 WM 80 60 40 20 50 100 150 WH 50 100 200 300 100 400 500 RH 150 600 200 250 RM SVC 3.1 two-node max SVC 4.1 two-node max SVC 4.2 two-node max 40 20 50 100 60 70/30/50 150

The left-hand curves used 168 RAID-5 disks and 4 fibre-channel host attachments.
9 2008 IBM Corporation

IBM Systems and Technology Group

Performance for OLTP workload


30 25 Response Time (ms) 20 15 10 5 0 0 50000 100000 150000 200000 250000 300000 Throughput (I/O per s) 2 Nodes 4 Nodes 6 Nodes 8 Nodes

1536 15K RPM disks configured as RAID-10; 32 fibre channels to host.


10 2008 IBM Corporation

IBM Systems and Technology Group

Scalability
300000 Max Throughput (Under 30 ms) 250000 200000 150000 100000 50000 0 2 4 SVC Nodes SVC 4.2 SVC 4.1 SVC 3.1 SVC 1.2.1 SVC 1.1.0 6 8

11

2008 IBM Corporation

IBM Systems and Technology Group

Sequential Throughput (Maximum)


12 10 Sequential MB/s Thousands 8 6 4 2 0 Read 3.1 Read 4.1 Read 4.2 4 Node Write 3.1 6 Node Write 4.1 8 Node Write 4.2

2 Node

12

2008 IBM Corporation

IBM Systems and Technology Group

SPC-2 Results through 3/1/2008 (SPC-2 MBPS)


0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000

IBM DS3400 Sun StorageTek 2530 (RAID-5) Sun StorageTek 2530 (mirrored) IBM SAN Volume Controller 4.2 IBM DS4800 Sun StorageTek 2540 (mirrored) Sun StorageTek 2540 (RAID-5) HP StorageWorks 1000 Fujitsu ETERNUS Model 1100 IBM DS4700 Model 70 Sun StorageTek 6140 (RAID-5) Sun StorageTek 6140 (mirrored) Sun StorageTek 6540 (RAID-5) Sun StorageTek 6540 (mirrored) Fujitsu ETERNUS Model 500 IBM DS4700 Model 70 IBM SAN Volume Controller 4.1 (4Gb) Sun 6140 (RAID-5) Sun 9985 Sun 6140 (Mirrored) Sun 3320 IBM-DS8300 Sun 3510 HP EVA 8000 Fujitsu ETERNUS Model 900 Sun 6130 IBM-SAN Volume Controller 3.1 (2Gb)

SVC is fastest

13

2008 IBM Corporation

IBM Systems and Technology Group

Software-Only Upgrade to SVC 4.2


40 35 Response Time (ms) 30 25 20 15 10 5 0 0 20000 40000 60000 80000 Throughput (I/O per s) 100000 120000

SVC 4.1 / 8F4 nodes SVC 4.2 / 8F4 nodes


70/30 Miss Workload

SVC 4.2 / 8G4 nodes

14

2008 IBM Corporation

IBM Systems and Technology Group

Bottom Line
SVC has proved it has the best performance in the storage industry.
More than 36 percent higher SPC-1 IOPS than any non-IBM product. More than double the SPC-2 data rate of any non-IBM product.

15

2008 IBM Corporation

IBM Systems and Technology Group

Space Efficient VDisks (Thin Provisioning)


Tests were performed with and without the use of FlashCopy. When using FlashCopy, COPYRATE=0 and CLEANRATE=0. The SE source storage was filled with null content before running the test workload (reflecting a database format operation). The SE target storage was *not* filled with null content.

16

2008 IBM Corporation

IBM Systems and Technology Group

Space Efficient Performance (2 nodes; OLTP)


30 25 Response Time (ms) 20 15 10 5 0 0 20000 40000 Throughput (I/O per s) Fully alloc VDisks
17

60000

80000

Space eff VDisks


2008 IBM Corporation

IBM Systems and Technology Group

Performance during FlashCopy (Fully Allocated Source)


30 25 Response Time (ms) 20 15 10 5 0 0 20000 40000 Throughput (I/O per s) 60000 80000

No FlashCopy Fully alloc target


18

Space eff target 4 space eff targets


2008 IBM Corporation

IBM Systems and Technology Group

Performance during FlashCopy (Space-Efficient Source)


30 25 Response Time (ms) 20 15 10 5 0 0 20000 40000 Throughput (I/O per s) 60000 80000

No FlashCopy Space eff target


19

4 space eff targets

2008 IBM Corporation

IBM Systems and Technology Group

Space Efficient VDisks -- Conclusions


When FlashCopy is not being performed, SE and Fully Allocated Vdisks compare well with each other in OLTP performance When running OLTP host work during FlashCopy, Fully Allocated target VDisks enjoy a significant performance advantage compared with SE target VDisks.

20

2008 IBM Corporation

IBM Systems and Technology Group

FlashCopy Background Data Rates


Except where noted, tests were performed with COPYRATE=100. The maximum background copy data rate of the system was strongly dependent on the state of DS4700 cache mirroring.

21

2008 IBM Corporation

IBM Systems and Technology Group

FlashCopy Background Data Rates (Best Achieved; DS4700 Cache Mirror disabled)
Background Copy Data Rate (GB/s) 6 5 4 3 2 1 0 00:00

01:00 Time (HH:MM)

02:00

03:00

8 nodes
22

6 nodes

4 nodes

2 nodes
2008 IBM Corporation

IBM Systems and Technology Group

FlashCopy Background Data Rates (DS4700 Cache Mirror Enabled)


Background Copy Data Rate (GB/s) 6 5 4 3 2 1 0 00:00

01:00

02:00

03:00 Time (HH:MM) 8 nodes

04:00

05:00

06:00

23

2008 IBM Corporation

IBM Systems and Technology Group

Effect of COPYRATE
Background Copy Data Rate (GB/s) 3
COPYRATE <--- =100 ---> =90

=80

=70

1
=60 =50 =40

0 00:00 01:00 Time (HH:MM)

=30

=20

=10

=0

02:00

24

2008 IBM Corporation

IBM Systems and Technology Group

FlashCopy Demonstration
50 Terabyte example database. Cache mirroring enabled. Heavy host load running to source (100,000 IOPS). Response time kept below 10 milliseconds by selecting COPYRATE. Repeated FlashCopies taken every hour.

25

2008 IBM Corporation

IBM Systems and Technology Group

Effect of COPYRATE Under Host Load


Background Copy Data Rate (GB/s) 3
COPYRATE <--- =100 ---> =90

40 Host I/O Response Time (ms)

=80

30

2 20
=70

1
=60 =50 =40 =30

10
=20 =10 =0

0 00:00 01:00 Time (HH:MM) Background Copy Host I/O (100,000 IOPS)

0 02:00

26

2008 IBM Corporation

IBM Systems and Technology Group

Initial FlashCopy (COPYRATE=80)


Background Copy Data Rate (GB/s) 3 30 Host I/O Response Time (ms)

20

10

0 00:00 01:00 02:00 03:00 04:00 05:00 06:00 07:00 08:00 09:00 Time (HH:MM) Background Copy Host I/O (100,000 IOPS)

27

2008 IBM Corporation

IBM Systems and Technology Group

Incremental Flashes Every Hour (COPYRATE=80)


Background Copy Data Rate (GB/s) 3 30 Host I/O Response Time (ms)

20

10

0 00:00 01:00 Time (HH:MM) 02:00

0 03:00

Background Copy Host I/O (100,000 IOPS)


28 2008 IBM Corporation

IBM Systems and Technology Group

FlashCopy Demonstration -- Conclusion


A 50 terabyte database can be flashed
Once per hour While supporting a host OLTP workload of 100,000 I/Os per second. At a response time under 10 milliseconds.

29

2008 IBM Corporation

IBM Systems and Technology Group

Global Mirror with Distance


40 35 Response Time (ms) 30 25 20 15 10 5 0 0 10000 20000 30000 40000 Throughput (I/O per s) 1,000 km 3,000 km 50000 60000

50 km 300 km
70/30/50 workload
30

10,000 km 30,000 km

2008 IBM Corporation

IBM Systems and Technology Group

Metro Mirror versus Global Mirror


40 35 Response Time (ms) 30 25 20 15 10 5 0 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 Throughput (I/O per s) GM / 4 nodes / 50 km
70/30/50 workload
31 2008 IBM Corporation

MM / 4 nodes / 50 km

IBM Systems and Technology Group

The Bottom Line: Consolidate Storage with SVC


Desktops Workstations Workstations Handheld devices Desktops

1950s Server-Centric
Server Terminals

LAN LAN SAN Server Server Server Information

System Storage

SAN Storage Storage

Subsystems

1990s Network-Centric

21st Century Information-Centric Globally Integrated Enterprise


2008 IBM Corporation

32

IBM Systems and Technology Group

References
IBM System Storage SAN Volume Controller Version 4.2 Performance, Bruce McNutt and Barry Whyte, IBM w3 System Sales site (http://w3-1.ibm.com/sales/systems/ & use search). IBM System Storage SAN Volume Controller Version 4.1.1 Global Mirror Performance, Bruce McNutt and Barry Whyte, IBM w3 System sales site (http://w3-1.ibm.com/sales/systems/ & use search). IBM System Storage SAN Volume Controller Version 4.1.1 Metro Mirror Performance, Bruce McNutt and Barry Whyte, IBM w3 System sales site (http://w3-1.ibm.com/sales/systems/ & use search). IBM System Storage SAN Volume Controller Version 4.3 FlashCopy Performance, Bruce McNutt, IBM w3 System Sales site (http://w31.ibm.com/sales/systems/ & use search).

33

2008 IBM Corporation

IBM Training

Trademarks and Disclaimers


The following are trademarks of the International Business Machines Corporation in the United States and/or other countries. For a complete list of IBM Trademarks, see www.ibm.com/legal/copytrade.shtml: AS/400, e business(logo), DBE,DS8000, ESCON, eServer, FICON, IBM, IBM (logo), iSeries, MVS, OS/390, pSeries, RS/6000, S/30, VM/ESA, VSE/ESA, WebSphere, xSeries, z/OS, zSeries, z/VM, System i, System i5, System p, System p5, System x, System z, System z9, BladeCenter, FlashCopy The following are trademarks or registered trademarks of other companies: Java and all Java based trademarks and logos are trademarks of Sun Microsystems, Inc., in the United States and other countries or both Microsoft, Windows,Windows NT and the Windows logo are registered trademarks of Microsoft Corporation in the United States, other countries, or both. Intel, Intel logo, Intel Inside, Intel Inside logo, Intel Centrino, Intel Centrino logo, Celeron, Intel Xeon, Intel SpeedStep, Itanium, and Pentium are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries. UNIX is a registered trademark of The Open Group in the United States and other countries or both. Linux is a trademark of Linus Torvalds in the United States, other countries, or both. Other company, product, or service names may be trademarks or service marks of others. NOTES: Any performance data contained in this document was determined in a controlled environment. Actual results may vary significantly and are dependent on many factors including system hardware configuration and software design and configuration. Some measurements quoted in this document may have been made on development-level systems. There is no guarantee these measurements will be the same on generally-available systems. Users of this document should verify the applicable data for their specific environment. IBM hardware products are manufactured from new parts, or new and serviceable used parts. Regardless, our warranty terms apply. Information is provided AS IS without warranty of any kind. All customer examples cited or described in this presentation are presented as illustrations of the manner in which some customers have used IBM products and the results they may have achieved. Actual environmental costs and performance characteristics will vary depending on individual customer configurations and conditions. This publication was produced in the United States. IBM may not offer the products, services or features discussed in this document in other countries, and the information may be subject to change without notice. Consult your local IBM business contact for information on the product or services available in your area.

34

2008 IBM Corporation

IBM Training

Trademarks and Disclaimers (continued)


NOTES: All statements regarding IBM's future direction and intent are subject to change or withdrawal without notice, and represent goals and objectives only. Information about non-IBM products is obtained from the manufacturers of those products or their published announcements. IBM has not tested those products and cannot confirm the performance, compatibility, or any other claims related to non-IBM products. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products. Prices are suggested US list prices and are subject to change without notice. Starting price may not include a hard drive, operating system or other features. Contact your IBM representative or Business Partner for the most current pricing in your geography. Any proposed use of claims in this presentation outside of the United States must be reviewed by local IBM country counsel prior to such use.

The information could include technical inaccuracies or typographical errors. Changes are periodically made to the information herein; these changes will be incorporated in new editions of the publication. IBM may make improvements and/or changes in the product(s) and/or the program(s) described in this publication at any time without notice.

Any references in this information to non-IBM Web sites are provided for convenience only and do not in any manner serve as an endorsement of those Web sites. The materials at those Web sites are not part of the materials for this IBM product and use of those Web sites is at your own risk.

IBM makes no representation or warranty regarding third-party products or services including those designated as ServerProven, ClusterProven or BladeCenter Interoperability Program products. Support for these third-party (non-IBM) products is provided by non-IBM Manufacturers. IBM may have patents or pending patent applications covering subject matter in this document. The furnishing of this document does not give you any license to these patents. Send license inquires, in writing, to IBM Director of Licensing, IBM Corporation, New Castle Drive, Armonk, NY 10504-1785 USA.

35

2008 IBM Corporation

You might also like