Skip to Content

CE 7.1 EHP1 SP03 and CE 7.2 SP03: Performance Comparison/Analysis

Very often we get into a situation where we have to take a call whether upgrade is at all necessary considering the current state of our project. What happens if performance is the only concern?

We have been using CE 7.1.1 EHP1 SP03 from the very outset of our project. Later on when CE 7.2 arrived, we had to take a judicious decision whether migration to the newer release would at all be necessary and worthy as far as performance is concerned (Though we are well aware of the fact that CE 7.2 has come up with lots of new features). As NWDI track migration between tracks of different versions is always a major concern, we decided to go for CE 7.2 if and only if there will be a significant performance improvement. For that we carried out a series of testing of services which first update a BO in CE then would call a service for updating tables in ECC and then return a response back to CE. We took notes of two scenarios:

 

1. Time taken for executing the service provided by CE

2. The time period between ECC service call and getting response back from ECC.  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

                     CE 7.1.1 EHP1

 

 

                             CE 7.2

 

 

Number of Updates

 

 

Overall Execution time(in ms)

 

 

ECC Service Call time(in ms)

 

 

Number of Updates(Same data used for CE 7.1)

 

 

Overall Execution time(in ms)

 

 

ECC Service Call time(in ms)

 

 

1

 

 

353

 

 

178

 

 

1

 

 

330

 

 

161

 

 

1

 

 

578

 

 

271

 

 

1

 

 

746

 

 

265

 

 

5

 

 

1543

 

 

394

 

 

5

 

 

1862

 

 

367

 

 

5

 

 

1759

 

 

574

 

 

5

 

 

2502

 

 

365

 

 

5

 

 

1300

 

 

176

 

 

5

 

 

1693

 

 

262

 

 

5

 

 

999

 

 

178

 

 

5

 

 

1512

 

 

268

 

 

5

 

 

1369

 

 

377

 

 

5

 

 

1719

 

 

263

 

 

1

 

 

609

 

 

180

 

 

1

 

 

630

 

 

264

 

 

1

 

 

720

 

 

272

 

 

1

 

 

636

 

 

268

 

 

1

 

 

351

 

 

175

 

 

1

 

 

744

 

 

274

 

                        Test results of both CE 7.1 and CE 7.2

When we analyze the results we can find only two instances where overall execution time improved in CE 7.2 whereas we have six instances where ECC service execution time got improved. That’s really an interesting outcome of our testing. I am very eager to hear all of your comments on this.

3 Comments
You must be Logged on to comment or reply to a post.
  • Hi Sunit,

    As a member of the CE/BPM performance team I’m of course interested in your analysis. Can you shed some light on your test cases (single user, volume or load tests), and as well on the implementation details (service composition, CAF, JPA) of the service calls?

    regards, Nick

    • Hi Nick,

      Thanks for your reply.Yes, these are test cases for single user and data volume was ranging from one sales order line item to five sales order line item each of those having around 50 fields , I used  CAF application service and CAF database. Web service(CAF application service was exposed as WS) call was made for interaction between ECC and CE.

      regards,
      Sunit

      • Dear Sunit,

        I need your kind help as I am looking for a sample list of recommended parameters values to be inserted into my CE environment.The parameter are as per below. I am really hoping you could help me here.

        Baseline:-

        Server
        -verbose:gc

        -XX:SurvivorRatio=2 -XX:TargetSurvivorRatio=90

        -XX:+UseParNewGC

        -XX:+HandlePromotionFailure

        -XX:+PrintGCTimeStamps

        -XX:+PrintGCDetails

        -XX:+UseTLAB

        -XX:SoftRefLRUPolicyMSPerMB=1

        -XX:ReservedCodeCacheSize=64M

        -XX:CodeCacheMinimumFreeSpace=2M

        -Djava.awt.headless=true

        -Djava.security.policy=./java.policy

        -Djco.jarm=1

        -Dorg.omg.CORBA.ORBClass=com.sap.engine.system.ORBProxy

        -Dorg.omg.CORBA.ORBSingletonClass=com.sap.engine.system.ORBSingletonProxy

        -Dorg.omg.PortableInterceptor.ORBInitializerClass.com.sap.engine.services.ts.jts.ots.PortableInterceptor.JTSInitializer

        -Djavax.rmi.CORBA.PortableRemoteObjectClass=com.sap.engine.system.PortableRemoteObjectProxy

        -Dsun.io.useCanonCaches=false

        -Djava.security.egd=file:/dev/urandom

        Disptacher
        -verbose:gc

        -Djava.security.policy=./java.policy

        Memory:-

        -Xmx4096m

        -Xms4096m

        -XX:NewSize=682m -XX:MaxNewSize=682m

        -XX:MaxPermSize=1024m -XX:PermSize=1024m

        Disptacher

        -Xms512M

        -Xmn70M

        -Xgcpolicy:gencon

        -verbose:gc

        – Xdump:system:events=systhrow,filter=java/lang/OutOfMemoryError,request=exclusive+prepwalk