by Raghuram Bharadwaj C
07/20/2004
Objective
The objective of this benchmark is to compare the performance of JMS vs. Web services for sequential request processing.
Assumptions
Assumptions
- Communication is synchronous (Client - Server)
- Java clients are sending string messages to the server
- Sequential Request Process -> stress test
Environment
The benchmark was tested under the following environment
- Java Web Services Developer Pack/1.2, JVM Version 1.4.1-b21, Apache/Tomcat Web Server
- MQ Series - Direct Connection, JDK 1.3.1/JDK1.4.1-b21
- Windows 2000 Version 5.0 Service Pack 4 - 512 MB RAM
- Windows 2000 Architecture - x86
Summary
The above results concluded that using Web services to connect to an external data source performs better than JMS under sequential request processing. If we go with JMS as an option, then MDB implementation would be a better choice.Note: The benchmarking is done based on above architecture. Please do not compare it with the benchmarking results that are available on the Internet.
Other possible options for performing the benchmark are as follows:
- Java clients sending XML messages to the server
- Concurrent Request Process - Stress test
- SOAP with attachments API
