Re: Which scales better? J2EE, PHP, or CGIDEV2?

Ahh.... now there's the trick is it not? Most of the benchmarks do not take real business work into account whereas the IBM CPW benchmark is all about real business load. If scientific workload is what is what's needed the other benchmarks work well. Real business workload, a different matter. Unfortuanatly we do not have a good web business benchmark to use right now. CPW does not really give a good indication of a web based workload although we can make some inferences from it.

Jim Oberholtzer
CEO/Chief Technical Architect
Agile Technology Architects, LLC


On 10/7/2010 10:07 AM, Thorbjoern Ravn Andersen wrote:

Den 06/10/10 20.13, Jim Oberholtzer skrev:

> While I don't ordinarily disagree with Paul, in this case I would
> agree with Kevin, the language has very little to do with scaleability.
> The server (both hardware and software ) has everything to with it.
> I'll put up a smaller IBM i/Zend Server against a larger Java server on
> any hardware most anytime and get equal or better results. Furthermore
> the Zend server is much easier to manage. Advanced configuration is
> about the same difficulty on both WebSphere and Zend servers so a toss
> up there.

The Jetty 7 java server has reportedly managed 100.000 concurrent comet
connections on a four core x86-server.

I don't know what they_did_ in each connection but I was impressed.
-- Thorbjørn Ravn Andersen "...plus... Tubular Bells!"