axis-java-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
Subject [jira] Commented: (AXIS-1497) AXIS2: performance improvements
Date Wed, 08 Sep 2004 11:09:44 GMT
The following comment has been added to this issue:

     Author: Venkat Reddy
    Created: Wed, 8 Sep 2004 4:08 AM
I' just had a look at the perf test results for Axis 1.2 at the following URL's as given in
Alek's post.

After going through the above, one will wonder why on earth people care about Axis? Just because
it is free and sounds homely because of Apache stamp on it? I think the performance should
the highest priority issue if we don't want these kinds of unpleasant reports.

Some observations from the above tests:
1. Axis is the SLOWEST engine of the group consisting of Sun's RI, JibxSoap, gSOAP, XSOAP4,
Axis and may be GLUE.
2. Axis's rpc style is unusable for a bit larger payloads and is not production ready.
3. It seems that AXIS-Java has huge memory leak - test was not completed as JVM ran out of
memory even though it was started with -Xmx1024m (1GB!) and it actually managed not only to
take all memory but also all swap space leading to machine freezing which is very bad sign
if you plans to run AXIS-Java based services for this kind of payloads.
4. Apache Axis framework was designed primarily to suit the needs of rpc/enc services, and
this legacy is going to limit its flexibility and performance going forward.
5. Java reflection used in the generic serializers adds to the performance issues.
6. DOM object creation even for doc/literal style consumes time.
7. Axis doesn't use custom parsers.

1. Which of observations listed above are valid?
2. Do these observations identify the problems correctly?
3. Can we translate the valid ones into specific bugs and tasks (could cover more than what's
in the observations listed above)?
4. Which one of these tasks/issues should go into Axis 1.2 release?

View this comment:

View the issue:

Here is an overview of the issue:
        Key: AXIS-1497
    Summary: AXIS2: performance improvements
       Type: Test

     Status: Unassigned
   Priority: Major

    Project: Axis
             Basic Architecture

   Reporter: Aleksander Slominski

    Created: Fri, 6 Aug 2004 10:29 AM
    Updated: Wed, 8 Sep 2004 4:08 AM
Environment: ALL

this is umbrella issue to record discussion and links to other issues related to performance,
memory footprint, and general architecture

at minimum i think it is important to profile current AXIS 1.x and identify places where performance
problems occur.

to facilitate this a set of tests/benchmarks should be used, currently i know about following

* SOAP Benchmark Suite for Scientific Computing 
-- source code available; that benchmark emphasizes support for very small and very large
messages (sending binary data and large arrays):

(i am main author of this benchmark so i can help with running it, interpreting results and
identifying problems/solutions)

* Dennis Sosnoski's Clean SOAP benchmark 
-- source code available; that benchmark seems to emphasizes doc/literal:

Additional links:

* "Performance and Load-Testing of Axis with Various Web Services Styles" by Rajal Shah, Naresh
Apte (Date: July 16, 2004)
--  no links to code used for testing?

This message is automatically generated by JIRA.

If you think it was sent incorrectly contact one of the administrators:

If you want more information on JIRA, or have a bug to report see:

View raw message