I would probably be looking at 2-4 machines with 4 gigs each to handle a load like this balanced with mod_backhand or lvs.
Ah, that's respectable :) With that amount of RAM you may have interest in looking into optimising parts of AxKit for your setup, such as playing with mmap here and there.
The majority of requests would be of the web service nature and not likely require a whole lot of transformation, in most cases serving raw xml, but with the need for transformation available.
It is very unlikely that transformation will ever be your bottleneck anyway, it's arguably the fastest part. Just how dynamic will your WS-style (REST?) request will be? How (from a distance) will they be constructed? It looks to me as though you might have a good use case for using a custom Provider, and that as your DOM will likely be synthetic you'll be interested in the latest get_dom() (currently only in CVS).
I am in the market for an app platform that doesn't have admitted memory leaks :o
Then one piece of advice (scarred by experience): pay very serious attention to the versions of libxml2, libxslt, XML::LibXML, and XML::LibXSLT you use. If you have a mix that works then stick to it until you *really* need to change. Those are high quality packages but they have their tough bits and here and there leaks are exposed that are triggered because of unpredicted interactions. Apart from that I don't remember seeing AxKit leak.
-- Robin Berjon <[EMAIL PROTECTED]> Research Engineer, Expway http://expway.fr/ 7FC0 6F5F D864 EFB8 08CE 8E74 58E6 D5DB 4889 2488
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
