Mark Foster wrote:
>
> Here's my understanding of how it works...
> Let's say I have 512mb RAM, and each of my apache+mod_perl processes uses an
> avg. of 8mb. Ignoring RAM needed for the system stuff (kernel, etc), I can
> hold a maximum 512/8=64 processes in memory. Anything more and I'm swapping
> to disk.
> Now with apache+jserv, if each process is about 2mb (prob. high), with an
> ApacheJserv process of about 24mb, I can hold (512-24)/2=244 processes.
> So an exlusive apache/jserv solution can serve 3.81x as many conncurrent
> connections.
> What am I missing?
Little things like processor power, application design, database speed,
query construction/optimization, SSL, application speed, speed of the
network, speed of remote data sources, and on, and on, and on.
The only thing you're referring to is how many processes you can fit
into memory - which has nothing to do with capacity.
If your design is poor/slow .. or if a query takes awhile to run .. or
you have lacking processor power .. or if you're dependant on remote
data sources...
.. you still won't be getting anywhere with 244 of those processes
sitting in memory. If anything, you'll REDUCE your performance by
filling memory with these processes, leaving no spare space for other
storage needs.
A few, well-designed, optimized processes will do MUCH more than many
poorly designed, inefficient ones will.
(note: I'm not flogging any particular language .. just your assumptions
:-)
-- dsp
--
--------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
READ THE FAQ!!!! <http://java.apache.org/faq/>
Archives and Other: <http://java.apache.org/main/mail.html/>
Problems?: [EMAIL PROTECTED]
