Benchmark » History » Revision 19
Revision 18 (chin-yeh, 07/07/2011 10:49 AM) → Revision 19/20 (chin-yeh, 07/07/2011 03:32 PM)
{{toc}}
h1. Benchmark
The purpose of this test is to benchmark the *throughout* and *response time* on a fully integrated system but the *asynchronous/background* processes are not included, e.g. queue consumer.
See issue #110 for the details on how to benchmark queue systems & web services.
h2. Test Coverage
!test_coverage.jpg!
h2. Component Used
|_.Name|_.Version|
|ecosway-commons|1.4-SNAPSHOT|
|ecosway-adaptor|1.2-SNAPSHOT|
|ecwyadaptor|0.5-SNAPSHOT|
|ecwyconnector|0.5-SNAPSHOT|
|ecwyqconnector-batch|1.4.1-SNAPSHOT|
|jms-adaptor|0.7-SNAPSHOT|
|soap-adaptor|0.6-SNAPSHOT|
|us-beans|1.4-SNAPSHOT|
|ws-mimic|2.6.1-SNAPSHOT|
|ws-status|1.3-SNAPSHOT|
h2. Tools
* JMeter
h2. Result
* Summary of this test - [attachment:summary.pdf]
* Appendix - [attachment:appendix.pdf]
When the total size of all messages is over the threshold, the queue system will just drop the new incoming message. It makes the queue system and message recipient(s) (e.g. web service) will not get overloaded.