Yellow, I've tested Zope and Apache on the SAME MACHINE Because of the Zope rendering on all objects before serving the request I've done a php script that reads an .html file and serves the output. Command used: ab -n 100 -c 10 <url> I didn't like the results!!!! RESULTS: APACHE Server Software: Apache/1.3.12 Server Hostname: 192.168.0.128 Server Port: 80 Document Path: /ger.php3 Document Length: 47349 bytes Concurrency Level: 10 Time taken for tests: 0.370 seconds Complete requests: 100 Failed requests: 0 Total transferred: 4751600 bytes HTML transferred: 4734900 bytes Requests per second: 270.27 Transfer rate: 12842.16 kb/s received Connnection Times (ms) min avg max Connect: 2 13 26 Processing: 29 22 14 Total: 31 35 40 ----------------------------------- ZOPE Server Software: Zope/Zope Server Hostname: 192.168.0.128 Server Port: 8080 Document Path: /rfc.html Document Length: 47349 bytes Concurrency Level: 10 Time taken for tests: 1.736 seconds Complete requests: 100 Failed requests: 0 Total transferred: 4802954 bytes HTML transferred: 4782249 bytes Requests per second: 57.60 Transfer rate: 2766.68 kb/s received Connnection Times (ms) min avg max Connect: 0 26 163 Processing: 42 137 130 Total: 42 163 293 ---------------------------------------- I think Zserver didn't actually give a very nice image of Zope!!! Any comments??????? --- Hugo Ramos - hramos@ruido-visual.pt Ruido Visual - Telecomunicacoes Interactivas, Lda
I think Zserver didn't actually give a very nice image of Zope!!! Any comments???????
Well, yeah, I don't think it's a very fair test. Of course Apache is going to be much faster in this sort of example, it's what it's good at: serving non-dynamic content up very fast. Even the PHP won't help much since I bet your PHP script didn't do authenticaion, aquistion, object serving, etc and I bet it didn't store that html file in a transactional, undo-able, object database. I see it as a pointless excercise since you're not comparing like for like. my $0.02 (tm) Chris
Hmmm... interesting results, but not quite as depressing as they may seem.
min avg max Apache/PHP total: 31 35 40 Zope (ZServer) total: 42 163 293
Speaking with my business head on, these figures are close enough that they don't matter. Zope's advantages as a development platform and low cost of ownership mean I would happily throw more hardware at a Zope solution to overcome any performance penalty that arises. And that's not even considering anything that ZEO will bring to the party. Generally we view hardware is pretty much free when compared to people, who are quite expensive. This breaks down on *really* high traffic sites, but there aren't that many of those around.... and if you find yourself there, there's always solutions like Squid and ZEO in your back pocket. I think I'll go back to being a developer now, before the rest of my group start asking for pay rises and/or new machines :-) Simon PS - don't forget Zope is fun to work on and an exciting community to be a member of - a major attraction. --------- My opinions are my own, NIP's opinions are theirs ---------- Simon J. Coles Email: simon@nipltd.com New Information Paradigms Work Phone: +44 1344 753703 http://www.nipltd.com/ Work Fax: +44 1344 753742 =============== Life is too precious to take seriously ===============
participants (3)
-
Chris Withers -
Hugo Ramos -
Simon Coles