IBM eServer pSeries
IBM
HC
server
R
U6
pSerie
s
IBM
HC
server
R U6
pSeries
IBM
HCR
server
U6
pSeries
IBM
HCR
U6
server
pSeries
July 13, 2004
© 2004 IBM Corporation
IBM eServer pSeries
Notes to Presenter
REQUIRED CHARTS
The presenter must display the Special Notices chart, the Notes on
Benchmarks and Values charts (if the referenced values are given), and
the Notes on Performance Estimates chart (if the referenced
performance estimates are given) during the course of the presentation.
Any printed copies of this presentation that are distributed must include
legible copies of these charts. If printed copies are not distributed, the
I to receive legible printed copies of
attendees must be offered the option
these charts.
TRADEMARKS
Please review the Special Notices page prior to updating this
presentation to ensure all trademarks used are given proper attribution.
SPEAKER NOTES
This presentation may contain speaker notes available imbedded or as a
separate file. Please ensure these are utilized if available.
Revised January 9, 2003
2
© 2004 IBM Corporation
IBM eServer pSeries
Notes to Presenter (Cont.)
To properly view this presentation, you will need to install the
IBM ~ True Type Font - IBMeserver. If it is not properly
installed, you will see IBM ^ or ~ instead of IBM ~.
The font can be downloaded by IBMers from:
I
http://w3.ibm.com/sales/systems/portal/_s.155/254?navID=f220s220t260&geoID=
All&prodID=pSeries&docID=eserverfont
or by IBM Business Partners from:
http://www.ibm.com/partnerworld/sales/systems; document: eserverfontbp
Revised May 6, 2004
3
© 2004 IBM Corporation
IBM eServer pSeries
Special Notices
This document was developed for IBM offerings in the United States as of the date of publication. IBM may not make these offerings available in other
countries, and the information is subject to change without notice. Consult your local IBM business contact for information on the IBM offerings available in
your area.
Information in this document concerning non-IBM products was obtained from the suppliers of these products or other public sources. Questions on the
capabilities of non-IBM products should be addressed to the suppliers of those products.
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.
All statements regarding IBM future direction and intent are subject to change or withdrawal without notice, and represent goals and objectives only.
The information contained in this document has not been submitted to any formal IBM test and is provided "AS IS" with no warranties or guarantees either
expressed or implied.
All examples cited or described in this document are presented as illustrations of the manner in which some IBM products can be used and the results that may
be achieved. Actual environmental costs and performance characteristics will vary depending on individual client configurations and conditions.
IBM Global Financing offerings are provided through IBM Credit Corporation in the United States and other IBM subsidiaries and divisions worldwide to
qualified commercial and government clients. Rates are based on a client's credit rating, financing terms, offering type, equipment type and options, and may
vary by country. Other restrictions may apply. Rates and offerings are subject to change, extension or withdrawal without notice.
IBM is not responsible for printing errors in this document that result in pricing or information inaccuracies.
All prices shown are IBM's United States suggested list prices and are subject to change without notice; reseller prices may vary.
IBM hardware products are manufactured from new parts, or new and serviceable used parts. Regardless, our warranty terms apply.
Many of the pSeries features described in this document are operating system dependent and may not be available on Linux. For more information, please
check: http://www.ibm.com/servers/eserver/pseries/linux/whitepapers/linux_pseries.html.
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. Some measurements quoted
in this document may have been estimated through extrapolation. Users of this document should verify the applicable data for their specific environment.
I
Revised February 6, 2004
4
© 2004 IBM Corporation
IBM eServer pSeries
Special Notices (Cont.)
The following terms are registered trademarks of International Business Machines Corporation in the United States and/or other countries: AIX, AIX/L, AIX/L(logo),
alphaWorks, AS/400, Blue Gene, Blue Lightning, C Set++, CICS, CICS/6000, CT/2, DataHub, DataJoiner, DB2, DEEP BLUE, developerWorks, DFDSM,
DirectTalk, DYNIX, DYNIX/ptx, e business(logo), e(logo)business, e(logo)server, Enterprise Storage Server, ESCON, FlashCopy, GDDM, IBM, IBM(logo),
ibm.com, IBM TotalStorage Proven, IntelliStation, IQ-Link, LANStreamer, LoadLeveler, Lotus, Lotus Notes, Lotusphere, Magstar, MediaStreamer, Micro Channel,
MQSeries, Net.Data, Netfinity, NetView, Network Station, Notes, NUMA-Q, Operating System/2, Operating System/400, OS/2, OS/390, OS/400, Parallel Sysplex,
PartnerLink, PartnerWorld, POWERparallel, PowerPC, PowerPC(logo), Predictive Failure Analysis, pSeries, PTX, ptx/ADMIN, RISC System/6000, RS/6000,
S/390, Scalable POWERparallel Systems, SecureWay, Sequent, ServerProven, SP1, SP2, SpaceBall, System/390, The Engines of e-business, THINK, ThinkPad,
Tivoli, Tivoli(logo), Tivoli Management Environment, Tivoli Ready(logo), TME, TotalStorage, TURBOWAYS, VisualAge, WebSphere, xSeries, z/OS, zSeries.
The following terms are trademarks of International Business Machines Corporation in the United States and/or other countries: AIX/L(logo), AIX 5L, AIX PVMe,
AS/400e, BladeCenter, Chipkill, Cloudscape, DB2 OLAP Server, DB2 Universal Database, DFDSM, DFSORT, Domino, e-business(logo), e-business on demand,
eServer, GigaProcessor, HACMP, HACMP/6000, i5/OS, IBMLink, IBM Virtualization Engine, IMS, Intelligent Micro-Partitioning, Miner, iSeries, NUMACenter,
POWER, POWER Hypervisor, Power Architecture, Power Everywhere, POWER Hypervisor, PowerPC Architecture, PowerPC 603, PowerPC 603e, PowerPC 604,
PowerPC 750, POWER2, POWER2 Architecture, POWER3, POWER4, POWER4+, POWER5, POWER5+, POWER6, Redbooks, Sequent (logo), SequentLINK,
Server Advantage, ServeRAID, Service Director, SmoothStart, SP, S/390 Parallel Enterprise Server, ThinkVision, Tivoli Enterprise, TME 10, TotalStorage Proven,
Ultramedia, VideoCharger, Visualization Data Explorer, X-Architecture, z/Architecture.
I
A full list of U.S. trademarks owned by IBM may be found at: http://www.ibm.com/legal/copytrade.shtml.
UNIX is a registered trademark in the United States and other countries licensed exclusively through The Open Group.
Linux is a registered trademark of Linus Torvalds in the United States, other countries or both.
Microsoft, Windows, Windows NT and the Windows logo are registered trademarks of Microsoft Corporation in the United States and/or other countries.
Intel, Itanium and Pentium are registered trademarks and Intel Xeon and MMX are trademarks of Intel Corporation in the United States and/or other countries
AMD Opteron ia a trademark of Advanced Micro Devices, Inc.
Java and all Java-based trademarks and logos are trademarks of Sun Microsystems, Inc. in the United States and/or other countries.
Other company, product and service names may be trademarks or service marks of others.
Revised July 1, 2004
5
© 2004 IBM Corporation
IBM eServer pSeries
Notes on Benchmarks and Values
The benchmarks and values shown herein were derived using particular, well configured, development-level computer systems. Unless otherwise indicated for a
system, the values were derived using external cache, if external cache is supported on the system. Buyers should consult other sources of information to evaluate
the performance of systems they are considering buying and should consider conducting application oriented testing. For additional information about the
benchmarks, values and systems tested, contact your local IBM office or IBM authorized reseller or access the following on the Web:
TPC
http://www.tpc.org
Linpack
http://www.netlib.no/netlib/benchmark/performance.ps
Pro/E
http://www.proe.com
SPEC
http://www.spec.org
GPC
http://www.spec.org/gpc
NotesBench Mail
http://www.notesbench.org
VolanoMark
http://www.volano.com
STREAM
http://www.cs.virginia.edu/stream/
Unless otherwise indicated for a system, the performance benchmarks were conducted using AIX V4.3 or AIX 5L. IBM C Set++ for AIX and IBM XL
FORTRAN for AIX with optimization were the compilers used in the benchmark tests. The preprocessors used in some benchmark tests include KAP 3.2 for
FORTRAN and KAP/C 1.4.2 from Kuck & Associates and VAST-2 v4.01X8 from Pacific-Sierra Research. The preprocessors were purchased separately from
these vendors. Other software packages like IBM ESSL for AIX and MASS for AIX were also used in some benchmarks.
The following SPEC and Linpack benchmarks reflect microprocessor, memory architecture, and compiler performance of the tested system (XX is either 95 or
2000):
–SPECintXX - SPEC component-level benchmark that measures integer performance. Result is the geometric mean of eight tests comprising the CINTXX
benchmark suite. All of these are written in the C language. SPECint_baseXX is the result of the same tests as CINTXX with a maximum of four compiler
flags that must be used in all eight tests.
–SPECint_rateXX - Geometric average of the eight SPEC rates from the SPEC integer tests (CINTXX). SPECint_base_rateXX is the result of the same tests
as CINTXX with a maximum of four compiler flags that must be used in all eight tests.
–SPECfpXX - SPEC component-level benchmark that measures floating-point performance. Result is the geometric mean of ten tests, all written in
FORTRAN, included in the CFPXX benchmark suite. SPECfp_baseXX is the result of the same tests as CFPXX with a maximum of four compiler flags
that must be used in all ten tests.
–SPECfp_rateXX - Geometric average of the ten SPEC rates from SPEC floating-point tests (CFPXX). SPECfp_base_rateXX is the result of the same tests as
CFPXX with a maximum of four compiler flags that must be used in all ten tests.
–SPECweb96 - Maximum number of Hypertext Transfer Protocol (HTTP) operations per second achieved on the SPECweb96 benchmark without significant
degradation of response time. The Web server software is ZEUS v.1.1 from Zeus Technology Ltd.
–SPECweb99 - Number of conforming, simultaneous connections the Web server can support using a predefined workload. The SPECweb99 test harness
emulates clients sending the HTTP requests in the workload over slow Internet connections to the Web server. The Web server software is Zeus from Zeus
Technology Ltd.
–SPECweb99_SSL - Number of conforming, simultaneous SSL encryption/decryption connections the Web server can support using a predefined workload.
The Web server software is Zeus from Zeus Technology Ltd.
–SPEC OMP2001 - Measures performance based on OpenMP applications.
–SPECsfs97_R1 - Measures speed and request-handling capabilities of NFS (network file server) computers.
I
Revised May 28, 2003
6
© 2004 IBM Corporation
IBM eServer pSeries
Notes on Benchmarks and Values (Cont.)
–SPECjAppServer200X
(where X is 1 or 2) - Measures the performance of Java Enterprise Application Servers using a subset of J2EE APIs in a complete endto-end Web application.
The Linpack benchmark measures floating-point performance of a system.
–Linpack DP (Double Precision) - n=100 is the array size. The results are measured in megaflops (MFLOPS).
–Linpack SP (Single Precision) - n=100 is the array size. The results are measured in MFLOPS.
–Linpack TPP (Toward Peak Performance) - n=1,000 is the array size. The results are measured in MFLOPS.
–Linpack HPC (Highly Parallel Computing) - solves the largest system of linear equations possible. The results are measured in GFLOPS.
STREAM is a simple synthetic benchmark program that measures sustainable memory bandwidth (in MB/s) and the corresponding computation rate for simple
vector kernels. Both standard and tuned results may be reported. http://www.cc.virginia.edu/stream/
VolanoMark is a 100% pure Java server benchmark that creates long-lasting network client connections in groups of 20 and measures how long it takes for the clients
to take turns broadcasting their messages to the group. The benchmark reports a score as the average number of messages transferred by the server per second.
–The
I
following Transaction Processing Performance Council (TPC) benchmarks reflect the performance of the microprocessor, memory subsystem, disk subsystem,
and some portions of the network:
–tpmC - TPC Benchmark C throughput measured as the average number of transactions processed per minute during a valid TPC-C configuration run of at least
twenty minutes.
–$/tpmC - TPC Benchmark C price/performance ratio reflects the estimated five year total cost of ownership for system hardware, software, and maintenance
and is determined by dividing such estimated total cost by the tpmC for the system.
–QppH is the power metric of TPC-H and is based on a geometric mean of the 17 TPC-H queries, the insert test, and the delete test. It measures the ability of the
system to give a single user the best possible response time by harnessing all available resources. QppH is scaled based on database size from 30GB to 10TB.
–QthH is the throughput metric of TPC-H and is a classical throughput measurement characterizing the ability of the system to support a multiuser workload in a
balanced way. A number of query users is chosen, each of which must execute the full set of 17 queries in a different order. In the background, there is an
update stream running a series of insert/delete operations. QthH is scaled based on the database size from 30GB to 10TB.
–$/QphH is the price/performance metric for the TPC-H benchmark where QphH is the geometric mean of QppH and QthH. The price is the five-year cost of
ownership for the tested configuration and includes maintenance and software support.
Revised February 6, 2004
7
© 2004 IBM Corporation
IBM eServer pSeries
Notes on Benchmarks and Values (Cont.)
The following graphics benchmarks reflect the performance of the microprocessor, memory subsystem, and graphics adapter:
–SPECxpc results - Xmark93 is the weighted geometric mean of 447 tests executed in the x11perf suite and is an indicator of 2D graphics performance in an X
environment. Larger values indicate better performance.
–SPECplb results (graPHIGS) - PLBwire93 and PLBsurf93 are geometric means of literal and optimized Picture Level Benchmark (PLB) tests for 3D
wireframe and 3D surface tests, respectively. Larger values indicate better performance.
–SPECopc results - Viewperf 7 (3dsmax-01, DRV-08, DX-07, Light-05, ProE-01, UGS-01) and Viewperf 6.1.2 (AWadvs-04, DRV-07, DX-06, Light-04,
medMCAD-01, ProCDRS-03) are weighted geometric means of individual viewset metrics. Larger values indicate better performance.
The following graphics benchmarks reflect the performance of the microprocessor, memory subsystem, graphics adapter and disk subsystem.
–SPECapc Pro/Engineer 2000i2 results - PROE2000I2_2000370 was developed by the SPECapc committee to measure UNIX and Windows workstations in a
comparable real-world environment. Larger numbers indicate better performance.
The NotesBench Mail workload simulates users reading and sending mail. A simulated user will execute a prescribed set of functions 4 times per hour and will
generate mail traffic about every 90 minutes. Performance metrics are:
–NotesMark - transactions/minute (TPM).
–NotesBench users - number of client (user) sessions being simulated by the NotesBench workload.
–$/NotesMark - ratio of total system cost divided by the NotesMark (TPM) achieved on the Mail workload.
–$/User - ratio of total system cost divided by the number of client sessions successfully simulated for the NotesBench Mail workload measured. Total system
cost is the price of the server under test to the client, including hardware, operating system, and Domino Server licenses.
I
Application Benchmarks
–SAP - Benchmark overview information: http:// www.sap-ag.de/solutions/technology/bench.htm; Benchmark White Paper September, 2000;
http://www.sap-ag.de/solutions/technology/pdf/50020428.pdf.
–PeopleSoft - To get information on PeopleSoft benchmarks, contact PeopleSoft directly or the PeopleSoft/IBM International Competency Center in
San Mateo, CA.
–Oracle Applications - Benchmark overview information: http://www.oracle.com/apps_benchmark/
–Baan - The Baan benchmark demonstrates the scalability of Baan ERP solutions. The test results provide the number of Baan Reference Users (BRUs) that
can be supported on a specific system. BRU is a single on-line user or a batch unit workload. These metrics are consistent with those used internally by both
IBM and Baan to size systems. To get more information on Baan benchmarks, go to http://www.ssaglobal.com.
–J.D. Edwards Applications - Product overview information at http://www.jdedwards.com.
Revised December 9, 2003
8
© 2004 IBM Corporation
IBM eServer pSeries
Notes on Performance Estimates
rPerf
–rPerf (Relative Performance) is an estimate of commercial processing performance relative to other pSeries systems. It is derived from an IBM analytical
model which uses characteristics from IBM internal workloads, TPC and SPEC benchmarks. The rPerf model is not intended to represent any specific
public benchmark results and should not be reasonably used in that way. The model simulates some of the system operations such as CPU, cache and
memory. However, the model does not simulate disk or network I/O operations.
–rPerf estimates are calculated based on systems with the latest levels of AIX 5L and other pertinent software at the time of system announcement. Actual
performance will vary based on application and configuration specifics. The IBM ~ pSeries 640 is the baseline reference system and has a value of 1.0.
Although rPerf may be used to approximate relative IBM UNIX commercial processing performance, actual system performance may vary and is
dependent upon many factors including system hardware configuration and software design and configuration.
–All performance estimates are provided "AS IS" and no warranties or guarantees are expressed or implied by IBM. Buyers should consult other sources
of information, including system benchmarks, and application sizing guides to evaluate the performance of a system they are considering buying. For
additional information about rPerf, contact your local IBM office or IBM authorized reseller.
I
Revised June 28, 2004
9
© 2004 IBM Corporation
IBM eServer pSeries
IBM ~® p5 570
Where doesI p5-570 fit?
What does it do?
10
© 2004 IBM Corporation
IBM eServer pSeries
IBM ~ p5 and pSeries® Product Lines
POWER4+™ and POWER5™
Servers
IBM
High-end
IBM
pSeries
server
Mid-range
IBM
pSeries 690
Entry deskside
IBM
pSeries 655
high-density
cluster
pSeries
IBM
pSeries
server
p5-570
I
p5-550
server
H C R U6
IBM
p5-520
pSeries 615
Model 6E3
p5-550
Cluster 1600
IBM
server
IBM
H C R U6
server
pSeries
IBM
IBM
p5-520
server
IBM
IBM
IBM
H C R U6
server
pSeries
IBM
pSeries 615
Model 6C3
11
Entry rack
I
B
M
© 2004 IBM Corporation
IBM
IBM eServer pSeries
Incredible Values and Performance at the
Mid-range of POWER5 Product Line…
the IBM ~ p5 570
HC
server
R U6
pSeries
HC
server
R U6
● Replaces the pSeries 650 and 670
mid-range products
pSeries
HC
server
R U6
pSeries
HC
server
R U6
pSeries
● Building block approach to help cover the mid-range server
opportunity from 2-, 4-way up to 16-way
symmetric multiprocessor
I
(SMP)
● Flexibility of virtualization, dynamic LPAR, on demand processing
● Retains mainframe inspired reliability, availability and serviceability
(RAS) capabilities from the POWER4+ product line
● Includes new POWER5 RAS enhancements for self-healing
and additional concurrency
● Outstanding value for a mid-range UNIX® server!
12
© 2004 IBM Corporation
IBM eServer pSeries
World Leadership Breakthrough Performance for
16-way SMP Transaction Processing
HC
server
R
U6
HC
server
R
U6
TPC-C benchmark results*:
809,144 tpmC
$4.95 / tpmC
#1
HC
server
R
U6
HC
server
R
U6
pSeri
es
pSeri
es
pSeri
es
pSeri
es
I
AIX 5L V5.3 with DB2® 8.1
Single SMP, non-clustered
16-way 1.9 GHz POWER5 processors
512GB of memory
13
*Note: Submitted 07/12/04, Availability 09/30/04
See www.tpc.org
© 2004 IBM Corporation
IBM eServer pSeries
Leadership Benchmark for SAP* 2-tier SD R/3 Users
(Using 1.9 GHz p5-570, 4-, 8- and 16-way)
6000
16-way
32-way
36-way
Scalability of p5-570
5000
5056
4 to 8-way = 98%
4900
5050
8 to 16-way = 94%
SD USERS
4000
I
3000
2600
2880
2000
1500
1313
1000
880
0
4-w
4-w HP
p5-570 1.5 GHZ
Itanium 2®
8-w
p5-570
8-w HP
1.5 GHZ
Itanium 2
16-w 16-w HP
p5-570 1.5 GHZ
Itanium 2
32-w NEC
36-w Sun
1.5 GHZ UltraSPARC IV
Itanium 2
* See http\\www.sap.com/benchmark (results as of 7/13/04)
14
© 2004 IBM Corporation
IBM eServer pSeries
Product Transition Path to IBM ~ p5 Servers
2Q2004
3Q2004
POWER4+
POWER5
TBD
p690
p5-570
16-way
p670
I
p5-570
12-way
TBD
p655
p5-570
8-way
15
p650
p5-570
2- or 4-way
p630
p5-550
2- or 4-way
p615
p5-520
2-way
© 2004 IBM Corporation
IBM eServer pSeries
Incredible Value at the Mid-range of POWER5
Product Line…the p5-570
IBM
HC
server
R
U6
pSerie
s
● Continuing to refresh and replace with
new technology
I
Added value
p5-570 POWER5
p650 POWER4+
Processor speed
1.5, 1.65, 1.9 GHz
1.2 or 1.45 GHz
512GB
64GB
up to 160
8 LPARs
Memory
Micro-Partitioning™*
I/O drawers
20
8
* optional feature
16
© 2004 IBM Corporation
IBM eServer pSeries
p5-570 Mid-range Target Opportunities
Provides leadership partitioning and performance, low
2- to 16-way price points and mainframe-inspired reliability
features for top mid-range UNIX opportunities
I
IBM
17
© 2004 IBM Corporation
IBM eServer pSeries
Financing Advantages*
Continuing in the footsteps of the powerful, reliable line of IBM UNIX servers –
the IBM ~ p5 570 will provide a comprehensive mid-range position into the
POWER5 product line. Complementing this exciting value, IBM Global Financing
can help make the difference.
IBM Global Financing’s portfolio of offerings and services designed specifically
for small and medium businesses, feature Istreamlined applications, simplified
contracts and fast approvals.
LINK TO: http://www-1.ibm.com/financing/it_products/hardware/servers/
pSeries.html
Financing from IBM Global Financing can make acquiring complete multivendor IT
solutions less expensive through competitive rates, fair market value leases,
innovative offerings, and special financing promotions.
Please visit our Web site for the latest on Special Promotions:
LINK TO: http://www-1.ibm.com/financing/promotions/index.html
For AP and EMEA go to ibm.com/financing and “select a country”
from the right navigation drop down list.
18
© 2004 IBM Corporation
*Note: this is for North American
clients only. Similar promotions
may be available in other geographies
IBM eServer pSeries
IBM ~ p5 570
Product Line
Capabilities
and
I
Evolutions
19
© 2004 IBM Corporation
IBM eServer pSeries
Product Transition Path to POWER5 ~ p5 Servers
2Q2004
3Q2004
POWER4+
POWER5
TBD
p690
p5-570
16-way
p670
I
p5-570
12-way
TBD
p655
p5-570
8-way
20
p650
p5-570
2- or 4-way
p630
p5-550
2- or 4-way
p615
p5-520
2-way
© 2004 IBM Corporation
IBM eServer pSeries
Homogeneous Environment Across the Infrastructure
to Help Lower TCO and Simplify Management
High-end
pSeries 690
IBM
Operational consistency
across all levels for the
system administrator
Mid-range
Large database server
IBM
I
IBM
IBM
pSeries 670
pSeries 650
p5-570
Application servers,
Web hosting
Entry
H C R U6
H C R U6
IBM
IBM
p5-550
IBM
server
p5-520
IBM
server
H C R U6
server
pSeries
IBM
21
IBM
server
H C R U6
server
H C R U6
server
pSeries
IBM
pSeries
IBM
p5-520
Department servers,
end-user applications
© 2004 IBM Corporation
IBM eServer pSeries
POWER4+ Technology
Across the Product Line
IBM
IBM
IBM
IBM
pSeries
server
IBM
Binary compatibility
from over $6M …
to under $6K!
Footprint,
packaging
1, 2 24-inch
frames
No. of SMPs
GHz clock
GB memory
Int. storage
PCI-X slots
I/O drawers
LPARs
Max. rPerf
Cluster 1600
HACMP™
8, 16, 24, 32
1.5, 1.7, 1.9
8 to 1024
18.7TB
10 to 160
8
32
104.17
Yes
Yes
22
1 24-inch
frame
4, 8, 16
1.5
4 to 256
7.0TB
10 to 60
3
16
46.79
Yes
Yes
I
1, 2 24-inch
frames
4, 8
1.7
4 to 64
2.6TB
3 to 23
1
4
21.87
Yes
Yes
p
S
e19-inch rack,
r deskside
i 1, 2, 4
e 1.2, 1.45
to 32
s 14.1TB
p
S
19-inch rack,
e
deskside
r
i 1, 2
1.2, 1.45 (2-w)
e 1 to 16
s 1.1TB
6 to 20 (6C4)
2 (6C4)
4 (6C4)
8.69
Yes (6C4)
Yes
6
N/A
N/A
4.41
Yes (6C3)
Yes
H C R U6
server
19-inch
rack
2, 4, 6, 8
1.2, 1.45
2 to 64
14.5TB
7 to 63
8
8
18.67
Yes
Yes
© 2004 IBM Corporation
IBM
IBM
server
H C R U6
server
pSeries
IBM
IBM eServer pSeries
POWER5 Technology
Product Evolution
IBM
IBM
p
S
e
19-inch rack,
r
deskside
i 2, 4
e1.65
s1 to 64
H C R U6
server
Footprint,
packaging
I
1, 2 24-inch
frames
1, 2 24-inch
frames
No. of SMPs
GHz clock
GB memory
Int. storage*
PCI-X slots*
I/O Drawers
LPARs
Max. rPerf
Cluster 1600
HACMP
23 *
Note: includes I/O drawer capacity; ** 2-, 4-, 8-way only
19-inch
rack
2, 4, 8, 12, 16
1.5**, 1.65, 1.9
2 to 512
38.7TB
6 to 163
20
160
77.45
Yes
Yes
IBM
15.2TB
5 to 60
8
40
19.66
Yes
Yes
© 2004 IBM Corporation
p
S
e rack,
19-inch
deskside
r
i 2
1.65
e 1 to 32
s 8.2TB
IBM
server
H C R U6
server
pSeries
IBM
6 to 34
4
20
9.86
Yes
Yes
IBM eServer pSeries
POWER5 Reliability
Across the Product Line
IBM
IBM
Mainframe-inspired RAS*
Reliability,
Availability and
I
Serviceability
Functions
*Note: Some RAS functions
**Note: Planned for 4Q04
may not be available at the
same time on SLES 9 and RHEL AS 3
24
p
p
S
S
Online firmware
updates**
e
e
First Failure Data
Capture
r
DDR ECC Chipkill memory r
i
Bit-steering/redundant
memory
i
e
Memory soft scrubbing
e
Redundant power,
s fans
s
Dynamic Processor Deallocation
H C R U6
server
IBM
IBM
server
H C R U6
server
pSeries
Deallocate PCI-X bus, L2/L3 cache
Persistent memory deallocation
Hot-plug PCI-X slots, fans, power
Internal LED diagnostics
Hot-swappable disk drives
© 2004 IBM Corporation
IBM
IBM eServer pSeries
IBM ~ Cluster 1600 Building Blocks*
IBM
High-end
IBM
IBM
Mid-range
IBM
pSeries
server
IBM
pSeries 670
pSeries 690
pSeries 655
p5-570
I
pSeries 650
IBM
server
H C R U6
server
pSeries
IBM
H C R U6
IBM
p5-550
pSeries 630
IBM
server
H C R U6
server
pSeries
IBM
IBM
server
H C R U6
server
pSeries
IBM
pSeries 615
Entry
25
p5-520
Manages POWER™ servers and
xSeries® servers from a single-point-ofcontrol.
© 2004 IBM Corporation
IBM eServer pSeries
IBM ~ p5 Model 570
POWER5 and
I
IBM Virtualization Engine™
systems technologies
26
© 2004 IBM Corporation
IBM eServer pSeries
What’s New with the p5-570
Shared L2
Distributed switch
28
Mem Ctl
● 1.5, 1.65 and 1.9 GHz 2-way POWER5
advanced technology processors
276 million transistors
Small, dense chips (0.13 microns)
I
389 sq/mm die size
1.5 to 1.9 GHz clock speeds
On-chip memory controller
L3 cache no longer in path to memory
© 2004 IBM Corporation
IBM eServer pSeries
p5-570 Server Peak Bandwidth per
30.4GB/sec x 8 = 243.2GB/sec for 16-way
1.9 GHz 2-way Chip
For 2-way chip
25.5GB/sec x 8 = 204.6GB/sec for 16-way
(using 533 MHz DDR2 memory)
4-way
L3
I
Ctlr
Ctlr
Shared L2
Shared L2
Distributed switch
Distributed switch
GX Bus
5.1GB/sec
Four I/O
Drawers
4GB/sec
29
M
E
M
O
R
Y
GX Bus on second processor card
5.1GB/sec (4-way only)
Six
PCI-X slots
I/O Hub
6
2nd I/O
Hub
Four I/O
Drawers
4GB/sec
4GB/sec
4GB/sec
I/O Hub adapter uses slot 6 real estate
© 2004 IBM Corporation
IBM eServer pSeries
Standard Virtualization Capabilities
1 CPU
2 CPUs
4 CPUs
4 CPUs
Linux
AIX 5L
V5.2
AIX 5L
V5.3
AIX 5L
V5.2
. . .
I
POWER Hypervisor™
● Includes:
- LPAR
- Dynamic LPAR (AIX 5L and SLES 9)
- Virtual LAN
30
© 2004 IBM Corporation
IBM eServer pSeries
Advanced POWER Virtualization Option
Partition Load Manager
AIX 5L
V5.3
AIX 5L
V5.2
Routing
Disks
Linux
AIX 5L
V5.2
pool of 20
Linux
4 CPUs
AIX 5L V5.3
4 CPUs
AIX 5L V5.3
AIX 5L v5.3
Virtual
IP
Virtual
Disks
2 CPUs
AIX 5L V5.3
2 CPUs
Hosting
Partition
I
Virtual I/O Paths
POWER Hypervisor
I/O
Storage
Network
LAN, WAN, ...
● This feature includes:
Feature no.
- Micro-Partitioning
- Shared processor pool
- Hosting LPAR for Virtual I/O (disk and LAN)
- Partition Load Manager for AIX 5L LPARs
31
© 2004 IBM Corporation
7940
7941
7942
Server
p5-520
p5-550
p5-570
IBM eServer pSeries
IBM POWER5 – Designed for Productivity
Efficient Utilization Helps to
LOWER COST of COMPUTING
POWER Hypervisor
Dynamic
Micro-Partitions*
Shared
Processor
Pools*
Virtual I/O
Virtual LAN*
Dynamic
LPAR
Capacity
Upgrade
on Demand*
(select models)
Simultaneous
Multithreading
I
●Allows the server to have more partitions than
processors, in which processing resources can be finetuned with allocations as small as 1/10th of a processor
and increments of 1/100th of a processor.
*Note: optional feature
32
© 2004 IBM Corporation
IBM eServer pSeries
IBM POWER5 – Designed for Productivity
Efficient Utilization Helps to
LOWER COST of COMPUTING
POWER Hypervisor
Dynamic
Micro-Partitions*
Shared
Processor
Pools*
Virtual I/O
Virtual LAN*
Dynamic
LPAR
Capacity
Upgrade
on Demand*
(select models)
Simultaneous
Multithreading)
I
●Provides a pool of processing power that is shared
between partitions helping to improve utilization and
throughput. Also can be changed dynamically to suit
ever-changing environments
*Note: optional feature
33
© 2004 IBM Corporation
IBM eServer pSeries
IBM POWER5 – Designed for Productivity
Efficient Utilization Helps to
LOWER COST of COMPUTING
POWER Hypervisor
Dynamic
Micro-Partitions*
Shared
Processor
Pools*
Virtual I/O
Virtual LAN*
Dynamic
LPAR**
Capacity
Upgrade
on Demand*
(select models)
Simultaneous
Multithreading
I
●Helps administrators reallocate system resources
between partitions and across operating systems,
including processors, memory, and I/O without rebooting
the partition or system for faster response to changing
workload requirements
*Note: optional feature
34
**Note: Not available on Linux RHEL AS 3
© 2004 IBM Corporation
IBM eServer pSeries
IBM POWER5 – Designed for Productivity
Efficient Utilization Helps to
LOWER COST of COMPUTING
POWER Hypervisor
Dynamic
Micro-Partitions*
Shared
Processor
Pools*
Virtual I/O
Virtual LAN*
Dynamic
LPAR
Capacity
Upgrade
on Demand*
(select models)
Simultaneous
Multithreading
I
●Supports physical sharing of disk drives,
communications adapters, helping to reduce the
number of expensive devices and improve system
utilization and administration.
●Allows high-speed, secure partition-to-partition
communications to help improve performance
*Note: optional feature
35
© 2004 IBM Corporation
IBM eServer pSeries
IBM POWER5 – Designed for Productivity
Efficient Utilization Helps to
LOWER COST of COMPUTING
POWER Hypervisor
Dynamic
Micro-Partitions*
Shared
Processor
Pools*
Virtual I/O
Virtual LAN*
Dynamic
LPAR
Capacity
Upgrade
on Demand*
(select models)
Simultaneous
Multithreading
I
●Servers with these features may be reconfigured
dynamically with additional processors or memory to
respond quickly to changing application workloads.
Helps clients pay for what they use when they use it.
With Reserve CoD the system can automatically adjust
processing power for LPARs assigned to the shared
processor pool.
*Note: optional feature
36
© 2004 IBM Corporation
IBM eServer pSeries
IBM POWER5 – Designed for Productivity
Efficient Utilization Helps to
LOWER COST of COMPUTING
POWER Hypervisor
Dynamic
Micro-Partitions*
Shared
Processor
Pools*
Virtual I/O
Virtual LAN*
Dynamic
LPAR
Capacity
Upgrade
on Demand*
(select models)
Simultaneous
Multithreading
I
●Simultaneous multi-threading technology allows latency
times of one thread of instructions to be utilized by another
totally independent thread of instructions, resulting in
improved system throughput. Each processor looks like
two to the system. Performance can be up to 30%** faster
than same speed systems without simultaneous multithreading.
*Note: optional feature; ** Based on IBM relative performance (rPerf) projections
37
© 2004 IBM Corporation
IBM eServer pSeries
POWER5 Micro-Partitioning for p5-570
Increases System Productivity
POWER4+
POWER5 SMP
SMP partitioning:
virtualization with
Micro-Partitioning*:
One partition per CPU
I
CPU 0
CPU 2
Dedicated,
Under-utilized
Resources
Multiple partitions
activated in each CPU
CPU 1
CPU 3
Up to 16 partitions
in a 16-way SMP
Up to 160
Micro-Partitions
in a 16-way SMP
*Note: optional feature
38
© 2004 IBM Corporation
IBM eServer pSeries
POWER5 Virtualization
AIX 5L V5.3 and Linux
POWER4+
Simplification through virtualization
● Micro-Partitioning*
AIX 5L V5.2
Flexibility
Dynamic LPAR
 Move resources between
partitions
Capacity on Demand
●
Shared processor pools*
●
Dynamic LPAR
●
Virtual I/O*, LAN
●
Partition Load Manager*
Reduced resources
● Fewer processors and I/O adapters
Helps to increase overall server throughput
I
AIX 5L V5.2
2 CPUs 4 CPUs
Linux
4 CPUs
AIX 5L AIX 5L
V5.2
V5.1
Disks
2 CPUs 4 CPUs 4 CPUs
AIX 5L AIX 5L
V5.2
V5.3
AIX 5L
V5.2
Routing
pool of 20
AIX 5L V5.3
Linux
AIX 5L V5.3
Linux
AIX 5L V5.3
8 CPUs
2 CPUs
Hosting
Partition
AIX 5L
v5.3
Virtual
IP
Virtual
Disks
Virtual I/O Paths
Hypervisor
POWER Hypervisor
Storage
I/O
Network
* Note: optional feature
39
© 2004 IBM Corporation
LAN, WAN, ...
IBM eServer pSeries
Partition Load Manager*
● Policy-based, automatic partition resource tuning
● Can dynamically adjust CPU and memory allocations
● Supports AIX 5L V5.3/V5.2 partitions
● p5-520, p5-550, and p5-570 systems
Before resource
tuning
CPU
CPU
Unbalanced
resource
allocation
CPU
Adjust
resource
allocation
based on
business
priorities
CPU
After tuning
CPU
I
Test
LPAR
CRM
LPAR
Finance
LPAR
Test
LPAR
CRM
LPAR
Finance
LPAR
3 CPU
5 CPU
6 CPU
1 CPU
3 CPU
10 CPU
Agent
Agent
Agent
Agent
Agent
Agent
* Note: optional feature
40
© 2004 IBM Corporation
IBM eServer pSeries
POWER5 Simultaneous Multi-threading
POWER5
(Simultaneous Multi-threaded)
Execution units utilization
POWER4+
(Single Threaded)
● Utilizes unused execution
System throughput
FX0
FX1
LS0
LS1
FP0
FP1
BRX
CRL
I
SMT
ST
unit cycles
● Presents SMP programming
model to software
● Natural fit with superscalar
out-of-order execution core
● Higher performance
Net result:
Higher system throughput
FX0
FX1
LS0
LS1
FP0
FP1
BRX
CRL
Legend
Thread0 active
No Thread active
Thread1 active
Execution units utilization
41
© 2004 IBM Corporation
IBM eServer pSeries
IBM ~ p5 570
I
Operating System
Options
42
© 2004 IBM Corporation
IBM eServer pSeries
POWER5 transitions on AIX 5L V5.2 and V5.3
AIX 5L V5.2
AIX 5L V5.3
POWER5 Hardware
POWER5 RAS functions
I
Simultaneous
Multi-threading
Dynamic LPAR
Dynamic LPAR
Advanced POWER Virtualization
Partition Load Manager
- Partition Load Manager
- Micro-Partitioning
- Shared processor pool
- Virtual I/O
New on demand capabilities
43
© 2004 IBM Corporation
IBM eServer pSeries
AIX 5L Software Delivers
ƒ Robust, scalable UNIX platform for
critical applications
ƒ Strong affinity with Linux for flexible
solutions that fit your business
ƒ Connections needed for e-business
and network computing
I
ƒ Security you can count on
ƒ Systems and network management that
puts you in control
ƒ Open and industry standards-based
platform that offers freedom of choice
ƒ Service and support to help keep a
business running
44
© 2004 IBM Corporation
IBM eServer pSeries
Linux on POWER*
● Enterprise Class Computing with Linux and POWER5
●
Flexibility with LPAR, DLPAR, and virtualization features
●
Reliability with built in self healing capabilities
●
Scalability from 2-way up to 16-way systems
● Linux distributions available for IBM ~ p5 are:
- SUSE LINUX Enterprise Server 9 for POWER (SLES 9)
- Red Hat Enterprise Linux AS 3 for POWER (RHEL AS 3)
I
● SUSE LINUX and Red Hat, Inc. also provide
support, upgrades and maintenance.
Technical support is also available through IBM SupportLine contract.
● Linux on pSeries remains orderable from IBM or directly from Linux distributors
● For more information about Linux running on IBM ~ p5 or pSeries systems:
http://www.ibm.com/servers/eserver/pseries/linux
http://www.redhat.com/software/rhel/as/
http://www.suse.com/us/business/products/server/sles/i_pseries.html
•Note, many of the features described in this document are operating system dependent and may not be available in Linux.
For more information, please check: www.ibm.com/servers/eserver/pseries/linux/whitepapers/linux_pseries.html.
45
© 2004 IBM Corporation
IBM eServer pSeries
IBM ~ p5 570
I
Product
Description
46
© 2004 IBM Corporation
IBM eServer pSeries
Building Block Modules Create SMP Mid-range Servers
● Configuration considerations of 16-way, 4-module server
M1
pSeries
Operator panel and service processor* required on first module.
If required, one or two HMCs attached to first module only.
Media required on first module only.
One disk drive required on first module only.
Attaches up to eight I/O drawers.
Plus four more I/O drawers for 12
M2
pSeries
I
Plus four more I/O drawers for 16
M3
pSeries
M4
pSeries
plus four more I/O drawers for a maximum of 20 I/O drawers
*Note: The service processor in the first or primary module
controls the two serial ports on the rear of that module.
If a HMC is chosen as an option, the two serial ports
will not be available for use. A PCI-X slot will be required
47
© 2004 IBM Corporation
IBM eServer pSeries
p5-570 SMP Building Block Modules Provide
Flexible Scaling Choices
4-way
8-way
HC
server
R U6
12-way
HC
server
R U6
HC
server
R U6
pSeries
1.5, 1.65 or 1.9 GHZ
2 to 128GB memory
6 to 61 PCI-X slots*
Up to 14.9TB storage*
Up to 40 LPARs
Up to 8 I/O drawers
16-way
HC
server
R U6
pSeries
pSeries
HC
server
R U6
HC
server
R U6
pSeries
HC
server
R U6
pSeries
pSeries
HC
server
R U6
pSeries
HC
server
R U6
1.5, 1.65 or 1.9 GHZ
8 to 256GB memory I
12 to 95 PCI-X slots* 1.65 or 1.9 GHZ
Up to 22.9TB storage* 12 to 384GB memory
Up to 80 LPARs
18 to 129 PCI-X slots* 1.65, or 1.9 GHz
Up to 12 I/O drawers Up to 30.8TB storage* 16 to 512GB memory**
Up to 120 LPARs
24 to 163 PCI-X slots*
Maximum of 20 RIO-2 I/O drawers per 16-way
Up to 16 I/O drawers Up to 38.7TB storage*
server and a maximum of four drawers
Up to 160 LPARs
per RIO-2 loop.
Up to 20 I/O drawers
*Note: Includes capacity from attached
7311-D20 I/O drawers.
48
pSeries
pSeries
HC
server
R U6
pSeries
**Note: 512GB is max for DDR1 memory,
64GB is max for DDR2 memory
© 2004 IBM Corporation
IBM eServer pSeries
Example of Eight I/O Drawers, Four I/O Hub Adapters
Optional
RIO-2 Hub
adapter
in slot # 6
off GX bus
16-way 1.9 GHz p5-570
Eight 7311 D20 I/O drawers
Standard
RIO-2 ports
off GX bus
Two I/O drawers per
4-way module attached
to two GX buses
4GB/sec
5.1GB/sec*
4GB/sec
5.1GB/sec*
4GB/sec
5.1GB/sec*
I
Leaves five available slots
in each 4-way module,
and 56 slots in the I/O
drawers for a total of
76 slots
This configuration
utilizes maximum
I/O bandwidth for a
1.9 GHz p5-570 server
4GB/sec
5.1GB/sec*
Total I/O peak bandwidth of 36.4GB/sec
* Note: Includes internal p5-570 slot bandwidth
49
Standard
RIO-2 ports
© 2004 IBM Corporation
IBM eServer pSeries
Building Block Modules Create SMP Mid-range Servers
● SMP flex cabling contains the processor fabric bus
● Up to 16-way server created using flat flex cabling to integrate
individual 4-way servers into a single SMP server (front view)
● Cabling may be detached to remove a module, and server
rebooted to operate with fewer resources
● SMP upgrades require new length cables
12-way
8-way
pSeries
I
16-way
pSeries
pSeries
pSeries
pSeries
pSeries
pSeries
pSeries
2-, 4-way
pSeries
pSeries
50
© 2004 IBM Corporation
IBM eServer pSeries
Building Block Modules Create SMP Mid-range Servers
● Service processor and clock use flat flex cabling to integrate
individual 4-way servers into a single SMP server (rear view)
8-way
12-way
16-way
I
2-, 4-way
51
© 2004 IBM Corporation
IBM eServer pSeries
Internal Structure of p5-570 Building Block Module
hot plug
power supply fans
service processor
Two hot-plug power
supplies
Six hot-plug, blind-swap PCI-X slots:
Five long 64-bit 133 MHz 3.3v
One short 64-bit 133 MHz 3.3v
Two hot-plug
I/O blowers
Two processor and memory
books, each with two
processors and up to 64GB of
memory. Processor options
include 1.5, 1.65, 1.9 GHz clock
speeds.
Maximum of 4-way and 128GB
DDR1 memory per module.
Minimum entry of 2-way and
2GB of memory.
Two processor
and memory books
Three hot-plug
processor power
regulators, one
optionally redundant.
Three required for
larger than 4-way
I
H C R U6
server
Two SMP flex cable
ports for interconnecting up to
four server building
block modules
(up to 16-way)
Hot plug
Op-panel required in
primary module
52
pSeries
Two hot-plug slimline media bays
(DVD-ROM standard)
The IBM 7212-102 1U media drawer
may be used for tape drives, etc.
RJ-45 connector
Six Ultra320 hot-swappable 10K rpm disks with
36.4GB, 73.4GB or 146.8GB disks and 15K rpm disks
with 36.4GB or 73.4GB capacities.
Maximum internal storage of 880.8GB
(dual SCSI controllers for 3+3 operation)
© 2004 IBM Corporation
IBM eServer pSeries
Rear View of p5-570 Building Block Module
Two SPCN
ports
2
3
4
5
S2
1
Rack
indicator
light port
6
Two serial
ports are active
on first module
only*
Two hot-plug AC power
supplies with dual
isolated line cords
I
S1
Two HMC
ports
System connect
Six hot-plug, blind-swap PCI-X slots:
for service processor
Five long 64-bit 133 MHz 3.3v
and clock, using
Slot #6 is short 64-bit 133 MHz 3.3v
Two 10/100/1000
flex cabling
Two USB 2.0
Mbps Ethernet
ports
ports
Two RIO-2 ports standard, two additional RIO-2 ports
*Note: If the HMC option is
optionally available via the Dual Port I/O Hub Adapter
chosen, the two serial ports
(not PCI-X) that pre-empts slot number 6
will not be available.
allowing up to four additional I/O drawers
53
© 2004 IBM Corporation
IBM eServer pSeries
p5-570 Supports I/O Drawer – 7311- D10
(Supported if migrated from another system, but not orderable as new)
Eight I/O drawers supported per 4-way module
Front View
28 inches deep
Two hot-plug,
redundant power
supplies for each
I/O drawer
4U high
I
Two I/O drawers fit side-by-side in 19-inch rack
Rear View
Rack
indicator
port
5V
Five 64-bit PCI-X hot-plug
slots and one PCI slot with
blind-swap capability.
(Five 133 MHz @ 3.3v
one 33 MHz @ 5v)
5V
Two SPCN
ports per
I/O drawer
PCI-X bus recovery at slot
level keeps system
running if a slot fails
Two RIO-2 ports per I/O drawer
54
© 2004 IBM Corporation
IBM eServer pSeries
p5-570 Optional I/O Drawer – 7311- D11
New model replacing the model D10 with full PCI-X
Eight I/O drawers supported per 4-way module
Front View
28 inches deep
Two hot-plug,
redundant power
supplies for each
I/O drawer
4U high
I
Two I/O drawers fit side-by-side in 19-inch rack
Rear View
Six 64-bit 133 MHz @ 3.3v
PCI-X hot-plug slots with
blind-swap capability.
Rack
indicator
port
PCI-X bus recovery at slot
level keeps system
running if a slot fails
Two SPCN
ports per
I/O drawer
Two RIO-2 ports per I/O drawer
55
© 2004 IBM Corporation
IBM eServer pSeries
p5-570 Optional RIO-2 I/O Drawer - 7311-D20
Ultra320 SCSI speed
(Up to eight supported*)
Front
Six-pack of
hot-swappable
disks; 36.4,
73.4 or 146.8GB
Op panel
Six-pack of
hot-swappable
disks; 36.4,
73.4 or 146.8GB
I
Rear
Two hot-plug
power
supplies
and fans. The
second is
optional
and redundant
110/220v AC
Seven PCI-X
64-bit 133 MHz
3.3v hot-plug
slots
not used
SPCN SPCN
Rack indicator light port
RIO-2 ports
*Note: Slot #6 of the p5-570 preempted if using more than four I/O drawers per module
56
© 2004 IBM Corporation
IBM eServer pSeries
p5-570 1.5, 1.65 or 1.9 GHz Processors
Book #2
Slot # 1
Shared L2
Shared L2
Book #1
Distributed switch
Slot
Slot
Slot
Slot
Shared L2
HCR
U6
server
Distributed switch
Mem Ctl
Distributed switch
pSeries
I
Slot
Slot
Slot
Slot
Slot
Slot
# 1Slot
#2
#3
# 4 Slot
Slot
Slot
# 5 Slot
#6
#7
#8
#2
#3
#4
#5
#6
#7
#8
Connector to backplane
● Processor cards protected by sturdy book packaging
● Books each have two POWER5 64-bit processors for a 4-way maximum
● Processors have 1.9MB L2 ECC shared cache and 36MB L3 ECC cache
● Separate service processor directs RAS activities on server
● Eight memory DIMMs are included on each card
● On demand processors and memory for peak workloads
● 1.65 GHz processors use DDR1 memory
● 1.5 GHz processors provide 8-way max, no DDR2 memory or CoD functions
57
© 2004 IBM Corporation
IBM eServer pSeries
p5-570 Capacity Upgrade On Demand - Processors
Slot # 1
Book #2
Shared L2
Shared L2
Distributed switch
Shared L2
Distributed switch
HCR
U6
server
pSeries
I
Slot
Slot
Slot
Distributed switch
Mem Ctl
Book #1
Slot
Slot
Slot
Slot
Slot
Slot
Slot
# 1Slot
#2
#3
# 4 Slot
Slot
Slot
# 5 Slot
#6
#7
#8
Connector to backplane
● On demand processors available to order for peak workloads
- A minimum of two processors with two activations is required.
- A maximum of 16 processors may be ordered with a choice of activations
depending on the reserve capacity required
● Dynamic Processor Sparing
- Inactive processors may be activated to replace a deallocated processor
58
© 2004 IBM Corporation
#2
#3
#4
#5
#6
#7
#8
IBM eServer pSeries
p5-570 Memory Quad Positioning
● Main memory ranges from 2GB to 128GB ECC using 266 MHz DDR1 SDRAM
per 4-way module
● Main memory ranges from 4GB to 16GB ECC using 533 MHz DDR2 SDRAM
per 4-way module. DDR2 memory only available on 1.9GHz processors using
1GB DIMMs only
● Memory is configured in two quads on each processor card,
and is pre-balanced by design for optimal performance with
each quad split between two synchronous memory interface (SMI) controllers.
● Plan ahead to reach desired maximumI memory to avoid discarding memory.
SMI
Shared L2
Distributed switch
Quad 1
Quad 2
SMI
*Note: DDR2 memory restricted to 1GB DIMMs
59
© 2004 IBM Corporation
Mem Ctl
2GB quad = 4 X 512MB DIMMs
4GB quad = 4 X 1GB DIMMs*
8GB quad = 4 X 2GB DIMMs
16GB quad = 4 X 4GB DIMMs
32GB quad = 4 X 8GB DIMMs
IBM eServer pSeries
p5-570 DDR ECC Chipkill Memory
Book #2 (4-way)
Book #2
Book #1
Slot # 1
Shared L2
Shared L2
Distributed switch
Slot
Slot
Slot
Slot
Shared L2
HCR
U6
server
Distributed switch
Mem Ctl
Distributed switch
pSeries
● Memory options of
2GB, 4GB, 8GB, 16GB, 32GB, 64GB or 128GB
Additional granularity is available between
binary sizes, e.g. 6GB, 10GB, 12GB, etc.
I
● 1GB DIMMs only for 533 MHz DDR2 memory
● CUoD memory not available for 533 MHz memory
● DIMMs are installed in quads
(four quads per server)
● Memory should be balanced between
the two processor books for best performance
60
Slot
Slot
Slot
Slot
Slot
Slot
# 1Slot
#2
#3
# 4 Slot
Slot
Slot
# 5 Slot
#6
#7
#8
#2
#3
#4
#5
#6
#7
#8
Book #1 (2-way)
Connector to backplane
● Eight memory DIMMs plug into slots
on each processor book
● DIMM sizes are: 512MB, 1GB, 2GB, 4GB
and 8GB for DDR1
(1GB DIMMs only for DDR2)
● 266 MHz memory may not be mixed with
533 MHz memory
● RAS features include ECC, Chipkill,
bit-steering to redundant memory bits,
and soft error scrubbing
© 2004 IBM Corporation
IBM eServer pSeries
p5-570 Capacity on Demand - Memory
Slot # 1
Book #2
Shared L2
Shared L2
Distributed switch
Shared L2
Distributed switch
HCR
U6
server
pSeries
I
Connector to backplane
● On demand memory available to order for peak workloads
(available on DDR1 266 MHz memory only)
● On demand memory not available on 1.5 GHz processor option
● Available feature is:
8GB
ordered
4GB Activated
● Maximum CoD memory is 32GB with 16GB active for a 4-way module
61
Slot
Slot
Slot
Distributed switch
Mem Ctl
Book #1
Slot
© 2004 IBM Corporation
Slot
Slot
Slot
Slot
Slot
Slot
# 1Slot
#2
#3
# 4 Slot
Slot
Slot
# 5 Slot
#6
#7
#8
#2
#3
#4
#5
#6
#7
#8
IBM eServer pSeries
2-way Model 570 Memory DIMM Positioning
Slot # 1
This set of examples is for a 2-way
which can have a memory range
of 2GB to 64GB* using one book
Slot # 2
Slot # 3
Slot # 4
Shared L2
Shared L2
Distributed switch
Slot
Slot
Slot
Slot
Distributed switch
#5
#6
#7
#8
Book #1 (2-way)
Connector to backplane
I
2GB
Book #1
4GB
Book #1
8GB
Book #1
32GB*
Book #1
16GB*
Book #1
64GB*
Book #1
DIMM Sizes
8GB
4GB
2GB
1GB
512MB
* Note: 16GB, 32GB and 64GB available only on 266 MHz DDR1 memory, 8GB is maximum for 533 MHz DDR2 memory for 2-way
62
© 2004 IBM Corporation
IBM eServer pSeries
4-way Model 570 Memory DIMM Positioning
For a 4-way, memory should be
balanced between both processor books
for best performance.
Slot # 1
Slot
Shared L2
Slot
Slot
Slot
Shared L2
Distributed switch
Book #1
Book #2
Shared L2
Distributed switch
Mem Ctl
Distributed switch
4GB
Slot
Slot
Slot
Slot
Slot
Slot
# 1Slot
#2
#3
# 4 Slot
Slot
Slot
# 5 Slot
#6
#7
#8
#2
#3
#4
#5
#6
#7
#8
Books #1 and #2 (4-way)
8GB
Book #1
Book #2
I
Connector to backplane
64GB*
16GB
Book #1
Book #1
Book #2
DIMM Sizes
Book #2
8GB
4GB
128GB*
32GB*
Book #1
Book #1
Book #2
Book #2
2GB
1GB
512MB
* Note: 32, 64 and 128GB available only on 266 MHz memory, 16GB is maximum for 533 MHz memory
63
© 2004 IBM Corporation
IBM eServer pSeries
Planning Ahead for Proper Memory Management
Slot # 1
Shared L2
Shared L2
Distributed switch
Slot
Slot
Slot
Slot
Shared L2
Distributed switch
Mem Ctl
Distributed switch
Slot
Slot
Slot
Slot
Slot
Slot
# 1Slot
#2
#3
# 4 Slot
Slot
Slot
# 5 Slot
#6
#7
#8
#2
#3
#4
#5
#6
#7
#8
Connector to backplane
I
● Plan ahead to avoid memory discarding
when upgrading:
- Determine where maximum desired memory will be and work
backwards to determine proper starting point
- If eventually growing to 128GB, start with a 32GB 266 MHz quad
2-way server, since it takes all DIMM positions on a 4-way to make 128GB
- If starting with a 4-way server and performance is critical, then
place a 32GB quad in each processor book
● Memory balancing is an important consideration for a 4-way server
- Better performance is achieved by balancing the quads
between the two processor books (highly application dependent)
64
© 2004 IBM Corporation
IBM eServer pSeries
Model 570 rPerf Performance Measurements
1.9 GHz processors
2-way = 11.16
4-way = 22.26
8-way = 42.14
12-way = 60.50
16-way = 77.45
Measured
using AIX 5Lv5.3
HC
server
R U6
pSeries
HC
server
R U6
pSeries
HC
server
R U6
pSeries
I
1.65 GHz processors
2-way = 9.86
4-way = 19.66
8-way = 37.22
12-way = 53.43
16-way = 68.40
HC
server
R U6
pSeries
1.5 GHz processors
2-way = 9.13
4-way = 18.20
8-way = 34.46
65
© 2004 IBM Corporation
IBM eServer pSeries
Hardware Management Console (HMC)
Redundant HMC
option
Ethernet
HC
server
R
U6
I
 System control software
- Focal point for virtualization management
- Cluster control point - Controls multiple non-clustered servers
(dependent on server type and mixture)
66
© 2004 IBM Corporation
IBM
pSerie
s
IBM eServer pSeries
Model 570s in T42 Rack Enclosure
p5-570 installs in T00, T42
and other 19-inch
non-IBM racks
IBM
Needs to be installed
below 36U point for
shipping
pSerie
s
pSerie
s
I
pSerie
s
pSerie
s
pSerie
s
pSerie
s
pSerie
s
pSerie
s
67
© 2004 IBM Corporation
IBM eServer pSeries
IBM 7316-TF3 Flat Panel Display Console
Rack-mounted flat panel console consisting of:
- IBM Flat Panel Color Monitor
- Flat Panel Monitor Rack-mounted Kit
- Rack Keyboard Tray
- IBM Space Saver 2 Keyboard which
mounts in Rack Keyboard Tray
(TrackPoint® mouse is integral to keyboard)
- Fits into a standard 19" rack with only 1U space
I
- Supported via GXT135P graphics adapter
- VGA switch supports up to eight systems
- Cascade eight switches to support a total of 64 systems
7 foot, 12 foot or 20 foot cable features
68
© 2004 IBM Corporation
HC
server
R
U6
pSeri
es
HC
server
R
U6
pSeri
es
HC
server
R
U6
pSeri
es
IBM eServer pSeries
TotalStorage Networks and Attachments
I
TotalStorage®
Products
69
© 2004 IBM Corporation
IBM eServer pSeries
IBM TotalStorage and Tivoli Ready to Go
Tape Storage
IBM
Disk Storage
LTO, Enterprise, External
IBM
ESS, FAStT, NAS
pSeries
IBM
server
p690
I
p655
H C R U6
IBM
server
IBM
H C R U6
server
pSeries
IBM
p5-570
p5-550
p5-520
SAN Switch
Storage
Virtualization
IBM, McDATA, CNT, Cisco
Tivoli® Storage
Software
70
© 2004 IBM Corporation
IBM eServer pSeries
IBM TotalStorage Ready to Go
AIX 5L V5.2/V5.3 and Linux
FAStT
ESS
IBM
IBM
pSeries
FAStT100
FAStT200
FAStT500
FAStT600
FAStT600 Turbo
FAStT700
FAStT900
Data
Retention
IBM
server
p690
I
H C R U6
IBM
server
IBM
H C R U6
server
ESS750
ESS800
p655
pSeries
IBM
p5-570
p5-550
p5-520
SAN Integration
Server
IBM
IBM
NAS
SCSI/SSA
H C R U6
IBM
DR450
71
2104-SCSI
7133-SSA
Gateway 500
2146
© 2004 IBM Corporation
IBM eServer pSeries
IBM TotalStorage Storage Area Network
Ready to Go
AIX 5L V5.2/V5.3 and Linux
CNT (Inrange)
IBM
IBM
FC9000 - 64 Director
FC9000 - 128 Director
FC9000 - 256 Director
IBM
pSeries
IBM
server
M12 & M14 Director
F08 & F16 SAN Switch
F32 SAN Switch
H08 & H16 SAN Switch
McDATA
p690
I
p655
H C R U6
IBM
server
IBM
H C R U6
server
pSeries
IBM
p5-570
p5-550
p5-520
Cisco
6064 & 6140 Director
2031 & 3232 SAN Switch
4300 & 4500 SAN Switch
FC-512 FibreCenter
72
MDS 9506 & 9509 Director
MDS 9120 & 9140 SAN Switch
MDS 9216 SAN Switch
© 2004 IBM Corporation
IBM eServer pSeries
IBM TotalStorage Tape Ready to Go
AIX 5L V5.2/V5.3 and Linux
IBM
LTO
Enterprise
IBM
IBM
pSeries
IBM
server
3580 Bridge box
3581 Autoloader
3582 Library
3583 Library
3584 Library
p690
I
p655
H C R U6
IBM
server
H C R U6
server
p5-570
3592 Drive
3590 drive
3494 Library
3494 VTS
p5-550
IBM
pSeries
IBM
p5-520
External Tape
DDS, DAT72, VXA-2
8mm, DLT, SDLT, SLR
73
© 2004 IBM Corporation
IBM eServer pSeries
IBM TotalStorage and Tivoli Software
Ready to Go
AIX 5L V5.2/V5.3 and Linux
IBM
TotalStorage
SAN Volume
Controller
Tivoli Storage
Manager
IBM
pSeries
IBM
server
p690
I
p655
H C R U6
IBM
TotalStorage
SAN Volume Controller
Blade
for Cisco MDS 9000
server
IBM
H C R U6
server
pSeries
IBM
Tivoli
Resource
Manager
p5-550
p5-520
TotalStorage
SAN File System
74
p5-570
Tivoli
SANergy
© 2004 IBM Corporation
Tivoli
SAN Manager
IBM eServer pSeries
Input/Output and Network Requirements
I
I/O and Network
Products
75
© 2004 IBM Corporation
IBM eServer pSeries
IBM Networking and I/O Ready to Go
Networking
Storage
Ethernet, Token-Ring
ATM, WAN
Fibre Channel, SCSI,
SCSI RAID, SSA
IBM
IBM
pSeries
IBM
server
Channels
I
ESCON, HiPPi
p690
p655
HC R
U6
IBM
server
HC R
U6
server
p5-570
p5-550
IBM
pSeries
IBM
Direct Talk
p5-520
Optical Media
DVD ROM, RAM & Tape
Encryption
DASD
10K and 15K rpm
76
Async
© 2004 IBM Corporation
IBM eServer pSeries
IBM Networking and I/O Ready to Go
AIX 5L V5.2/V5.3 and Linux
Networking
Ethernet
10/100 Ethernet
Quad Port 10/100
Gigabit Ethernet (GigE)
Dual port GigE
10GigE*
Token-Ring
WAN
2 Port
4 Port
Storage adapters
IBM
IBM
pSeries
IBM
server
I
p655
HC R
U6
IBM
server
HC R
U6
server
IBM
pSeries
IBM
p5-570
8 Port
128 Port
p5-550
Direct Talk
Optical Media
DVD ROM
DVD RAM
DDS4 Tape
VXA2 Tape
8mm Tape
DAT72* Tape
DASD
10K Ultra320 SCSI
36, 73 and 146GB
15K Ultra320 SCSI
36GB and 73GB
DTXA-2
Encryption
FIPS4 Coprocessor
SSL accelerator
*Note: new adapters
77
Async
p5-520
Channels
ESCON
HiPPi
p690
2GB FCAL
Dual Port
Ultra320 SCSI
Dual Port
U320 SCSI RAID
SSA
© 2004 IBM Corporation
IBM eServer pSeries
I/O Adapter Enhancements
10 Gigabit Ethernet-SR PCI-X Adapter
Provides connection on standard
shortwave multimode optical cable.
-Distances up to 33m using 62.5 um
multimode fiber
-Distances up to 66m with
50 um
I
multimode fiber
DAT72 4mm 36/72GB tape
Provides higher capacity 4mm tape storage
78
© 2004 IBM Corporation
IBM eServer pSeries
IBM ~ p5 570
Reliability
I
Availability and
Serviceability
79
© 2004 IBM Corporation
IBM eServer pSeries
Reliability Availability Serviceability (RAS)
Online
updating
Keep
thefirmware
system operational!
Online parts replacement
Utilize system redundancies
I
Deallocate system resources if necessary
Resolve intermittent faults with no intervention
Diagnose failures if they happen
Fault avoidance - designed to avoid downtime
80
© 2004 IBM Corporation
IBM eServer pSeries
Reliability Availability Serviceability*
RAS enhancements available on POWER5 servers
● Applications remain operational while IBM system
firmware is updated for most operations
(planned for 4Q04)
● ECC has been extended toI inter-chip connections
for the Fabric/processor busses (data, address, control)
● Partial L2 cache deallocation
● L3 cache line deletes improved from two to ten for better
self-healing capability
*Note: some RAS functions may not be available at the same time on Linux SLES 9 and RHEL 3.3
81
© 2004 IBM Corporation
IBM eServer pSeries
A Self-diagnosing Server
● First Failure Data Capture is key to a self-diagnosing system
and is an IBM exclusive!
● Critical parts of system can provide real-time
machine status information if faults occur
● Error data is continually being monitored by system
I
in order to take appropriate actions
● Many errors can be self-corrected or functions varied offline
while server remains available
● Server has capability to determine which part or component
needs repair and can phone IBM Service to provide precise
parts for maintenance at a time acceptable to client
82
© 2004 IBM Corporation
IBM eServer pSeries
Reliability Availability Serviceability
24x7
I
83
© 2004 IBM Corporation
IBM eServer pSeries
Capture Error Causes in Real-time
Cornerstone that is basis
of pSeries RAS technology
● First Failure Data Capture provides
error information in real-time
and makes
I
it possible to determine parts
necessary to fix problem
● Trying to recreate intermittent errors at a later
time can be extremely time consuming,
ineffective, and sometimes impossible
84
© 2004 IBM Corporation
IBM eServer pSeries
First Failure Data Capture Technology
 Old Failure Recreate Strategy
Error Checkers
CPU
Fault Isolation Register (FIR)
(unique fingerprint of each
error captured)
L1 Cache
L2/L3 Cache
Service
Processor
–Run diagnostic test cases during service call
–Repair based on test case symptom
–Questionable correlation to original problem
–Development focus on improved test cases
–Open service action plan if failure not recreated
–Test cases used to try to isolate failures and verify
correct operation
Log Error
Memory
Nonvolatile
RAM
Disk
Reliably identify each failing component,
reducing costly downtime
85
I First Failure Data Capture: an IBM Exclusive
–Specialized hardware designed to capture failure
data at time of failure
–Repair based on root-cause analysis
–Direct correlation to original client problem
–Engineering focus on built-in error detection and
capture
–Service action plan driven by captured failure
information
–Test cases used to only to verify correct operation
© 2004 IBM Corporation
IBM eServer pSeries
Start Off With Quality-based Design
Designing extra quality into
system to keep errors from
ever happening
I
● Mainframe-inspired components and technologies
- Designed for reduced power consumption,
cooler operating temperatures and increased reliability
86
© 2004 IBM Corporation
IBM eServer pSeries
An Active Server Immune System
● Bit-steering to redundant memory in event of failed memory module
to keep server operational
I
● Single bit error correction via ECC without reaching error thresholds
for main, L2 and L3 cache memories
● L3 cache line deletes extended from two to ten for additional self
healing
● ECC extended to inter-chip connections on fabric/processor busses
● Soft-error scrubbing to help prevent memory faults
87
© 2004 IBM Corporation
IBM eServer pSeries
ECC Memory Integrity and Availability
Memory scrubbing for soft singlebit errors that are corrected in the
background while memory is idle,
to help prevent multiple-bit errors.
Dynamically reassign memory I/O
Failing memory bit steered via bit-steering if error threshold
to spare memory bit
is reached on same bit
Chipkill
X
X
....
Spare
memory bits
I
Scatter memory chip bits across four separate ECC words for Chipkill recovery
Bit-scattering allows normal single bit ECC error processing, thereby
keeping system running with a Chipkill failure.
Bit-steering allows memory lines from spare memory bits to be
dynamically reassigned to a memory module with a faulty line to
keep system running.
If all spare bits are utilized, and threshold is reached, the Service Processor
will be invoked to request deferred maintenance at a time acceptable to
client.
88
© 2004 IBM Corporation
IBM eServer pSeries
Built-in Redundancies Designed for
Non-stop Operation
Use of redundancy to remain
operational with full resources
I
● Redundant spare memory bits
● Redundant fans
- Fans will speed up to try to compensate for failed fan
● Redundant power supplies (optional)
89
© 2004 IBM Corporation
IBM eServer pSeries
Continued Operation and System Integrity
Corrections and retries that
don't exceed threshold limits
and don't require client or
IBM CE intervention
I
● CEC bus retry and recovery
● PCI-X bus recovery
● ECC Chipkill soft errors
90
© 2004 IBM Corporation
IBM eServer pSeries
Keep Running With Fewer Resources
If threshold reached, resources can be
deallocated with system remaining
operational, allowing client to
plan maintenance at a convenient time
I
● Partial L2 cache deallocation
● PCI-X bus deallocation to the card level
● Removal and replacement of failed PCI-X cards
● Upon reboot, failed components will be deconfigured
to avoid repeat of errors
91
© 2004 IBM Corporation
IBM eServer pSeries
PCI Bus and Card Error Recoveries
PCI-X
Controller
X
I
● Problem with damaged connection results in PCI-X bus error
- That particular slot is varied offline by system
- All other slots remain active
- System stays operational without the slot active
- If PCI card is damaged, replace with new one
- If slot is damaged, plan for maintenance at client’s convenience
92
© 2004 IBM Corporation
IBM eServer pSeries
Replacement of Parts While System Runs
Online hot-plug of defective
electromechanical components
keeps system operational
● Disk drives
I
● Cooling fans
● Power subsystems
● PCI, PCI-X adapter cards
93
© 2004 IBM Corporation
IBM eServer pSeries
Quicker Service for Client Convenience
Increasing service productivity means
more client uptime
I
● Most firmware updates may be made while the server
remains operational (planned for 4Q04)
● Internal LED diagnostics identify components that
require service
● LEDs on I/O provide status of PCI-X slots and
disk drives for power, hot-swap and need for service
94
© 2004 IBM Corporation
IBM eServer pSeries
Health Care of p5-570 System
Service Processor
Processor Drawer
Processors
Memory
Power
Thermal
Fans
AIX Error Log
Central
repository
for all
system
errors
Hot Standby Backup
In unlikely event of a
unresolvable system
fault, system can be
backed up by
another
hot-standby system
95
Runs asynchronously.
Cycles continuously
Does not degrade
looking for faults
performance.
Service
Focal Point
HC
server
R
U6
Input/Output
AIX 5L device
drivers provide error
information toI
AIX 5L error log
AIX 5L Service Agent
Sends latest error
information to CE
and/or client when
necessary
High Availability
software: HACMP
HAGEO, GeoRM
Service Dispatch
Initiates
maintenance
action with parts
required to fix real
problem
© 2004 IBM Corporation
IBM
pSerie
s
IBM eServer pSeries
IBM ~ p5 570
Convenient Starting
Points for
I
Value Pak
Configurations
Makes Ordering Easy
96
© 2004 IBM Corporation
IBM eServer pSeries
IBM ~ p5-570 Value Paks
HC
server
R
U6
HC
server
R
U6
pSerie
s
HC
server
R
U6
HC
server
R
U6
pSerie
s
pSerie
s
pSerie
s
Model 570
I
 Offering popular, easy-to-order configurations
at considerable savings
 Choice of Linux, AIX 5L or no operating system
 Allows clients to modify these base configurations
to match up with specific application requirements
97
© 2004 IBM Corporation
IBM eServer pSeries
p5-570 1.5 GHz Express Value Paks
HC
server
R U6
pSeries
HC
server
R U6
pSeries
p5-570 (1.5 GHz)
Two processors, 4GB memory
one no-charge activation,
two 73.4GB disks
 p5-570 (1.5 GHz)
Four processors, 8GB memory,
two no-charge activations,
two 73.4GB disks
 p5-570 (1.5 GHz)
Eight processors, 16GB DDR1 memory,
four no-charge activations,
two 73.4GB disks

98
I
© 2004 IBM Corporation
IBM eServer pSeries
p5-570 2-, 4- and 8-way Value Paks
HC
server
R U6
pSeries

p5-570 (1.65 GHz)
Two processors, 4GB memory,
one no-charge activation,
two 73.4GB disks
p5-570 (1.65 GHz)
Four processors, 8GB memory,
two no-charge activations,
two 73.4GB disks
 p5-570 (1.9 GHz)
Four processors, 8GB DDR1 memory,
two no-charge activations,
two 73.4GB disks
 p5-570 (1.9 GHz)
Four processors, 8GB DDR2 memory,
two no-charge activations,
two 73.4GB disks

99
HC
server
R U6
pSeries
I
p5-570 (1.65 GHz)
Eight processors, 16GB DDR1 memory,
four no-charge activations,
two 73.4GB disks
 p5-570 (1.9 GHz)
Eight processors, 16GB DDR1 memory,
four no-charge activations,
two 73.4GB disks
 p5- 570 (1.9 GHz)
Eight processors, 16GB DDR2 memory,
four no-charge activations,
two 73.4GB disks

© 2004 IBM Corporation
IBM eServer pSeries
p5-570 16-way Value Paks
HC
server
R U6
pSeries
HC
server
R U6
pSeries
HC
server
R U6
pSeries


100
p5-570 (1.9 GHz)
Sixteen processors, 32GB DDR1 memory,
eight no-charge activations,
two 73.4GB disks
p5-570 (1. 9 GHz)
Sixteen processors, 32GB DDR2 memory,
eight no-charge activations,
two 73.4GB disks
I
HC
server
R U6
pSeries
© 2004 IBM Corporation