News: Performance Testing your Distributed J2EE Applications
In this article, Frank Teti describes a reusable mechanism for capturing application-bound, performance statistics for highly distributed J2EE applications. He examines the design and edits required to collect basic instrumentation/timing estimates for a sample Internet Order Entry Application distributed across a 5-tier architecture.
- Posted by: Nitin Bharti
- Posted on: February 16 2004 21:23 EST
Read Designing Performance Testing Metrics into Highly Distributed J2EE Applications
- I wish same hardware with other OS by Srinivas Chidumalla on February 20 2004 09:08 EST
- Performance Testing your Distributed J2EE Applications by Nick Xidis on March 03 2004 12:57 EST
- interesting approach by Jack Bruce on March 11 2004 14:04 EST
- JMS compliance by Frank Teti on March 12 2004 09:09 EST
- Bit behind the curve here, but... by william c on March 16 2008 20:37 EDT
It would be interesting to see same hardware configuration with Linux.
All fine with results, but with all those windows code floating around on the internet, i think that does not creates a much interest :)
Nice to see good technical content. Appreciate the details.
interesting approach... should be extended for how you can do web client through business and integration tiers performance testing, though. But I guess that would involve significant refactoring the HTML client side (JSP or XSLT process).
Also, not sure if Candle really has a tool that could reproduce the instrumentation functionality you designed for here.
Like the UML stuff. Good to see people actually use that stuff
Lastly, the JMS compliance of Seebeyond with respect to the have JMS messages has the product been upgraded to support message type?
actually this model was designed and used well over a year ago. I would think that the vendor would probably be completly compliant by now.
As far as Candle, I have all the respect in the world for that venerable vendor of system management software -- but they are expensive.
be cool, Frank
...System.currentTimeMillis() will limit your testing; you'll have to change the system clock to simulate timing issues.