> Hits per second is pretty low -- between 1 and 3 per server. The pages
> are very dynamic, though, and response time varies from almost
> instantaneous (1 second or less) up to 20, depending on the page and
> the current load. Some of these delays are expected to get no faster
> than 3-5 seconds turnaround, because of reliance on external data
> sources -- but they start piling up when you're serving a couple
> requests a second and the requests take a few seconds to complete.

Ok, let's see. We've got a few what I would consider medium-weight sites. All
database backed (local PostgreSQL server). They are all primarily
web/information sites, although we have few "web application" type situations.
In a few months we'll be finishing a very large scale web application running on
Webware/Postgres, but that doesn't help you now. Most of our sites are
PSP-based, but with all the logic held in Servlets that the PSPs inherit from.
Some are very dynamic, some less so.

Example one: a customer with a couple thousand product combinations, who gets a
medium amount of traffic, although it varies heavily based on day and hour.
Monthly average is about 70,000 page hits, but this is generally concentrated
into afternoons with 10-15 per minute. All with several sql queries / joins per
request. average 0.5 seconds per product page, from webware's Activity log.

Example two: also on this server, we have about 10 other webware-driven sites
that don't create as much load individually, but perhaps combined are 2-3 times
that load. This is all from a single webware instance, as contexts. This server
(1Ghz, 1GB mem Dell) shows no significant loading.

Example three: artificial load testing of our database environment. we've got
two test pages: one that makes a large join query and then uses python to cycle
through the results (several logical levels, several hundred results), and
another that makes several million select queries in a recursive loop. The first
consistently serves in 2-8 seconds, the second in 15-40 seconds. We've run our
test scripts with between 5 and 20 concurrent agents, making hundreds of requests.

Not terribly scientific (I'm running ApacheBench right now against our test
pages.... almost exactly the same general results), but perhaps some help.

Also, we've had very enjoyable success with our combination of
webware/apache/postgres, so we are currently in the middle of developing a
large-scale application (10K's total users, 100K's transactions/week... haven't
run more in-depth projections) using it. But this promise of large app
development doesn't really help your immediate need of examples.

- Luke

-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/


-------------------------------------------------------
This SF.net email is sponsored by: Etnus, makers of TotalView, The debugger 
for complex code. Debugging C/C++ programs can leave you feeling lost and 
disoriented. TotalView can help you find your way. Available on major UNIX 
and Linux platforms. Try it free. www.etnus.com
_______________________________________________
Webware-discuss mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/webware-discuss

Reply via email to