090606 Dell and SAP Benchmark-Portal-INFOBrief 2-legally approved

Description

   INFOBrief Dell - Your next ®SAP NetWeaver Portal-based Solution?   Results  • On April 7, 2006, SAP AG certified the results of a two‐tier SAP® Employee Self‐Service Portal (EP‐ESS) Standard Application Benchmark on a Dell PowerEdge 6850 server reaching a total of 1941 1SAP EP‐ESS Benchmark users . • Dell demonstrates its technology leadership in delivering the highest 2two‐tier SAP EP‐ESS Benchmark result as of May 4, 2006.   • The purpose of this INFOBrief is to share Dellʹs lessons learned in setting up a stable, long‐lasting SAP NetWeaver® Portal‐based solution.  Main steps that may assist in increasing performance and thus help to lower TCO can be found in this document.  Dell Product Group/Enterprise Page 1 April 2006 UPUP The Challenge and the Configuration  As the enterprise services architecture (ESA) gains momentum, more and more customers take the chance to integrate the SAP® NetWeaver Portal as a collaboration platform into their SAP NetWeaver platform‐based landscape. Because the Java‐based portal is quite different from ABAP‐based SAP® applications such as Sales and Distribution or Materials Management, research into optimal configuration settings was needed. The benchmark environment used is based on Novellʹs SUSE Linux Enterprise Server. Many customers with longstanding UNIX® backgrounds and investments are looking to Linux as a more cost‐effective alternative ...

Informations

• Publié par: Irkoi
• Nombre de lectures: 29
• Langue: Español

Extrait

INFOBrief Dell  Your next ® SAP NetWeaver Portalbased Solution?

Results •OnApril7,2006,SAPAGcertifiedtheresultsofatwo‐tierSAP® EmployeeSelf‐ServicePortal(EP‐ESS)StandardApplication BenchmarkonaDellPowerEdge6850serverreachingatotalof1941 1 SAPEP‐ESSBenchmarkusers . •Delldemonstratesitstechnologyleadershipindeliveringthehighest 2 two‐tierSAPEP‐ESSBenchmarkresultasofMay4,2006. •ThepurposeofthisINFOBriefistoshareDellʹslessonslearnedin settingupastable,long‐lastingSAPNetWeaver®Portal‐based solution.Mainstepsthatmayassistinincreasingperformanceand thushelptolowerTCOcanbefoundinthisdocument.

Dell Product Group/Enterprise

Page 1

April 2006

The Challenge and the Configuration U Astheenterpriseservicesarchitecture(ESA)gainsmomentum,moreand morecustomerstakethechancetointegratetheSAP®NetWeaverPortal asacollaborationplatformintotheirSAPNetWeaverplatform‐based landscape.BecausetheJava‐basedportalisquitedifferentfromABAP‐ basedSAP®applicationssuchasSalesandDistributionorMaterials Management,researchintooptimalconfigurationsettingswasneeded. ThebenchmarkenvironmentusedisbasedonNovellʹsSUSELinux EnterpriseServer.ManycustomerswithlongstandingUNIX® backgroundsandinvestmentsarelookingtoLinuxasamorecost‐effective alternative. Benchmarkdetails U Thebenchmark’sconfigurationandresultsare:

Hardware

Application Database Operating System SAP EPESS Benchmark users

Number of Top Level Navigation steps/hour

Dell Product Group/Enterprise

Dell™PowerEdge™6850 4processors/8cores/16threads IntelXeonMP3.0Ghz 32 GB RAM SAP NetWeaver Portal 6.0 MaxDB 7.5 3 Novell® SUSE™ Linux Enterprise Server 9 EM64T P P

1,941 583 000

Page 2

April 2006

What does it mean for your business? TheSAPEP‐ESSStandardApplicationBenchmarksimulatesreal‐lifeusers doingtypicalemployeeactivitieslikerecordingtheirworkingtimeor lookingattheirbankinformation.Theback‐endinfluenceonthe performancehasbeenomitted.Thegoalistoreachanaverageresponse timeoflessthantwosecondswithamaximumusernumber.Withthis usernumberinmind,youcancomparedifferenthardwareplatformsfor theneedsofyourbusinessplatform. Performance Parameters TherearemanyconfigurationparametersfortheSAP®NetWeaverPortal thatcanbefine‐tunedforperformance.Thechallengingtaskistofind thosethatreallymakeadifferencewithregardstobusinessneeds.For example,withUNIX/Linuxsystems,thereisoftenthenotionthatrunning agraphicinterfacecouldadverselyaffecttheoverallsystemperformance, soyoucouldgainperformancebydisablingit.Toverifythis,Dellranthe benchmarkoncewithafullyfeaturedKDEdesktopenvironmentrunning, andoncewithout.TherunwiththeKDEgraphicinterfaceshowedno significantchangesintheresponsetime.Thisshowsthat‐providedthe absenceofresourcebottlenecks‐theactivationofasystemʹsdesktop interfacewouldnottypicallyinterferewithitsportalsfunction. Thefollowingaresettingsthatthisbenchmarkshowedtobeof significance. BIOSSettings Thehardwareprefetcheroptimizesperformanceforsequentialmemory access.Wefoundanincreaseinperformanceifitisdisabled,whilethe PowerEdge6850’sdefaultsettingshaveitenabled.Toturnitoff,dothe following: −Whenbootingtheserver,entertheDellsystemBIOSbypressingF2 −InthesystemBIOS,choseCPUinformation‐>SequentialMemory Access −Setthisoptionto“disabled” Wefoundthattheprefetcherhadthefollowingperformanceimpactonthe SAPNetWeaverPortal‐basedsystem:

Dell Product Group/Enterprise

Page 3

April 2006

9

8 7 6

5 4

3 2 1

0

Prefetcher's influence

response time

OFF ON state Figure 1 Performance impact of thePrefetchersetting (lower is better) Figure1showsthatswitchingofftheprefetcherdecreased,inthisparticular example,thesystemʹsansweringtimefrom8.47secondsto2.16seconds, whileallotherparametersremainedunchanged.

Dell Product Group/Enterprise

Page 4

April 2006

SoftwareSettings ThesoftwaresettingsweredonefortheSAPJ2EEengineusingSAPʹs configtool.Itisagraphicconfigurationinterface:

Figure 2 Screenshot of theConfigtool Inthisexample,wefindtheJ2EEenginecontainingeightservernodes (serverIDnumbers3480750‐3480757).Theseventhnodehasbeenopened, itsexecutiontagisenabled,anditsheapconsumptioninthisexampleis 2,500MB.

Dell Product Group/Enterprise

Page 5

April 2006

ServerNodes U Oneofthedecidingperformancefactorsisthenumberofservernodes withintheSAPJ2EEengine.Theoptimalnumberiscontingenton availablememoryandCPUresources.ForaDell™PowerEdge6850with dual‐coreprocessorsand32GBRAM,wefoundsevenservernodes deliveroptimalperformancewhilethesystemʹsresponsetimesdifferas describedinthefollowingdiagram. Server Nodes' influence

2.25

2 1.75 1.5

1.25 1

0.75 0.5 0.25

response time

0 6 7 8 server nodes Figure 3 Performance impact of the server nodes' count(lower is better) Thediagraminfigurethreeshowstheminimumresponsetimeofthe examplesystematsevenservernodeswith1.92seconds,comparedto2.16 secondsforsixandeightservernodes.Foreachcolumn,thethreadcount forallservernodesremainedunchanged.

Dell Product Group/Enterprise

Page 6

April 2006

ThreadCount U Theapplicationthreadcountturnedouttoinfluencetheperformance significantly.Tomanipulatethethreadcount,usetheconfigtooltosetthe valuesforMinThreadCount,InitialThreadCountandMaxThreadCountunder cluster‐data‐>Globalserverconfiguration‐>managers‐> ApplicationThreadManager Threads' influence

2.25

2 1.75 1.5

1.25 1

0.75 0.5 0.25

response time

0 30 40 50 threads Figure 4 Performance impact of the threads' number(lower is better) Thediagraminfigurefourshowstheminimumresponsetimeofthe examplesystemat40threadsperservernode.Theresponsetimeis1.92 secondshere,comparedto1.94at30and2.02at50threadspernode.All otherparametersremainedunchanged. Bottom Line U TuningtherightparametersinanSAP®NetWeaverPortal‐basedsystem deliverssignificantperformanceincreases.Whileinourteststhehardware (BIOS)settingshaveyieldedthemostimportanteffect,thesoftware parametersplayanimportantroleinfinetuningthesystem,especially whenmappingthesoftwareconfigurationtothehardwareresources available.DistinctCPUormemoryresourcesneedadaptedsettingsin ordertobeoptimallyused.Findingtherightfithelpsyourbusiness platformtoruninamorestablemanner,deliveringthehighestpossible performanceandthusprovidingamoreoptimalreturnoninvestment.

Dell Product Group/Enterprise

Page 7

April 2006

1 TheDellresultonthetwo‐tierSAPEmployeeSelf‐ServicePortal(EP‐ESS)Standard P P ApplicationBenchmarkwascertifiedonApril7,2006,certificationnumber2006017,andfully complieswiththeguidelinesissuedbytheSAPBenchmarkCouncil.Moreinformationon SAPStandardApplicationBenchmarksisavailableathttp://www.sap.com/benchmark. 2 ForfulldetailsoftheSAPEP‐ESSStandardApplicationBenchmarkpleasereferto P P http://www.sap.com/solutions/benchmark/ep‐ess.epx. TH TH 3 Intheinterestofcustomerchoice,DellsupportstheLinuxdistributionsofbothNovell® P P SUSE™LinuxEnterpriseServer,aswellasRedHat®EnterpriseLinuxforallSAP implementations. ©2006DellInc.Dell,PowerEdge,andtheDelllogoaretrademarksofDellInc.Dell disclaimsanyproprietaryinterestinthemarksandnamesofothers.

Dell Product Group/Enterprise

Page 8

April 2006