[jbossts-issues] [JBoss JIRA] (JBTM-1369) Benchmark performance difference between JacORB and JDK ORB

Michael Musgrove (JIRA) issues at jboss.org
Tue Sep 2 09:51:01 EDT 2014


     [ https://issues.jboss.org/browse/JBTM-1369?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Michael Musgrove resolved JBTM-1369.
------------------------------------
    Resolution: Done


h2. Executive Summary:

Throughput for the sun orb is 2.5x faster than JacORB (1084 versus 453 distributed transactions/sec)

Slightly more detailed Summary:

h2. ORB transport comparison:

This test compares JTS performance using 2 different orbs (JacORB and the sun JDK orb).
The sun orb is the standard one bundled in the Oracle jdk and JacORB is the one we ship in product:

* sun orb: Java HotSpot(TM) 64-Bit Server VM (build 24.51-b03, mixed mode)
* jacorb: 2.3.2-jbossorg-5 (which I presume maps on to JacORB 2.3)

The test config runs a client and server in separate JVMs and uses explicit (OTS) transaction propagation to update a variable in a remote CORBA object (so the test uses the 1PC optimization).

The figures show that throughput for the sun orb is 2.5x faster than JacORB:
{quote}
2014-05-14 23:11:29,939 out: Test performance (for orb type JAVAIDL): 1084 calls/sec (200000 invocations using 50 threads with 0 errors. Total time 184399 ms)
2014-05-14 23:19:24,591 out: Test performance (for orb type JACORB): 453 calls/sec (200000 invocations using 50 threads with 0 errors. Total time 440744 ms) 
{quote}

h3. Remark: 
We also ran on a version of the IMB orb (IBM J9 VM build 2.6, JRE 1.7.0 Linux amd64-64 Compressed References 20140106_181350 (JIT enabled, AOT enabled)) but do note that we have not yet got recovery working with this orb so the following figure is provisional (IBM orb being 4x faster than JacORB):
{quote}
2014-05-14 22:43:50,588 out: Test performance (for orb type IBMORB): 1622 calls/sec (300000 invocations using 50 threads with 0 errors. Total time 184924 ms)
{quote}

Also noteworthy is that we have a job running the benchmark using narayana master.

> Benchmark performance difference between JacORB and JDK ORB
> -----------------------------------------------------------
>
>                 Key: JBTM-1369
>                 URL: https://issues.jboss.org/browse/JBTM-1369
>             Project: JBoss Transaction Manager
>          Issue Type: Task
>      Security Level: Public(Everyone can see) 
>          Components: JTS, Performance Testing
>    Affects Versions: 4.17.0
>            Reporter: Michael Musgrove
>            Assignee: Michael Musgrove
>             Fix For: 5.0.4
>
>
> JBTM-934 added support for the JDK orb to JTS. IOR sizes can range in size from 512 bytes to entirely unbounded, with the ORB being able to add arbitrary data within specific portions. Since the IOR is packed into transaction logs and if the IOR sizes are significantly different we may find that transaction throughput has been improved or degraded. We should run some benchmarks to see what the impact is.



--
This message was sent by Atlassian JIRA
(v6.3.1#6329)


More information about the jbossts-issues mailing list