This is archived documentation for an older version of Nirvana (v3.1). Please refer to documentation for the latest version if required.

Nirvana Benchmarks

Nirvana development and product evolution has continued to provide best of breed middleware messaging technology. The my-Channels technology development effort is founded on 3 themes, stability, scaleability and performance.

Many middleware vendors make outrageous claims about scaleability and performance. Within the JMS arena many vendors have published their own benchmark applications in an attempt to support their claims. In many instances these benchmarking results describe message throughput but do very little to describe what was happening on both the server and client machines while the benchmarks were running.

At my-Channels all benchmarking is done based on third party benchmarking software. my-Channels have adopted other vendor's "Try it out" offer and used their benchmark applications to benchmark against Nirvana.

The results underscore our commitment to stability, scaleability and performance.

SonicMQ Vs Nirvana

SonicMQ 3.0 and Nirvana 1.1 build 1466 were used in this test run. SonicSoftware's own benchmarking software was used to produce the results.

In all tests Nirvana performed significantly better than SonicMQ.

SonicMQ tests involving 25 and 50 connections.

Click here see full test results.
Other JMS Vendors Vs Nirvana

For legal reasons we are no longer in a position to provide performance comparisons with another well known JMS vendor. We encourage our customers to view this vendors results directly and to compare them against my-Channels*

*Please note that by clicking on this link you are leaving the my-Channels web site. (Terms of Use)

 

The Test Environment

Client Machine

Dell PowerEdge 1400SC . 2GB Ram. 2 x 1Ghz Intel PIII processors. 100BaseTX Network. Windows 2000 Server
Server Machine

Intel P4 1.7Ghz. 1GB Ram. 100BaseTX Network. Windows 2000 Workstation
Summary

In both comparisons above the server and client components of each product were used without modification or parameter configuration. The SonicMQ test harness program does not provide a mechanism to test message selectors so that could not be done.

All tests were using persistent topics and 1k messages.

If as a vendor or middleware provider you have a benchmarking application and would like my-Channels to use it in a performance comparison the please contact benchmark@my-Channels.com.

If as a client or user of middleware products you would like to understand how Nirvana continues to come out on top please contact info@my-Channels.com.