Re: [squid-users] Evaluating SQUID as a Reverse Proxy for a large web site

From: kapil khanna <[email protected]>
Date: Sun, 7 Mar 2004 13:13:43 -0600

Henrik,
I am using Apache as my web server. I decreased the dish cache size to 1 MB
and re-ran my tests. This time i got all TCP_MEM_HITS for the images and
static files. However my benchmark results were not better. The web server
still scaled much better.
My guess is that SQUID cannot handle concurrency too well. How do i get
SQUID to increase the no of processes or threads? Is my only option to have
many SQUID servers on different ports on a host with lot of RAM, running in
front of a load balancer to handle concurrency?
Thanks,
--Kapil
----- Original Message -----
From: "Henrik Nordstrom" <hno@squid-cache.org>
To: "kapil khanna" <kapil@j2eerules.com>
Cc: <squid-users@squid-cache.org>
Sent: Sunday, March 07, 2004 3:28 AM
Subject: Re: [squid-users] Evaluating SQUID as a Reverse Proxy for a large
web site

> On Sat, 6 Mar 2004, kapil khanna wrote:
>
> > I have been evaluating SQUID to deploy in front of a large web site to
cache
> > all static content (Images, JS Files, CSS Files, HTML files etc...) for
the
> > web site. I used JMeter as a load testing tool to evaluate the
scalability
> > of SQUID. This is my current config:-
> > cache_mem - 256MB
> > disk cache - 10MB.
> > I purposely have a very low disk cache so that i can get most out of
> > in-memory caching of static content. I also set content expiry (if not
set)
> > for images, JSP files etc to
> > 14400 80% 43200
>
> You probably should run without any disk cache at all in this
> configuration. If not Squid will not actually be able to use all that
> cache_mem..
>
> > Why is that the Web application scales better than SQUID? The one thing
> > that stands out is that SQUID is running as one process one thread,
> > whereas the web site is multithreaded.
>
> What kind of web server are you using?
>
> The benefits of using a cache infront of the web server is mostly seen if
> the web server can not handle very many concurrent connections. The cache
> then helps both by offloading the static content any by reusing the same
> persistent connections for multiple clients.
>
> Regards
> Henrik
>
>
Received on Sun Mar 07 2004 - 12:13:50 MST

This archive was generated by hypermail pre-2.1.9 : Thu Apr 01 2004 - 12:00:01 MST