[squid-users] Re: Recent Polygraph results for Squid?

Unfortunately numbers in that range is rare and usually not published.
In fact barely no one publishes and usable performance numbers about
their setups.

We try to collect such indications when given at
http://wiki.squid-cache.org/KnowledgeBase/Benchmarks but not very much
of value even there with lots of missing details....

The biggest Squid installation reasonably public about it's numbers is
Wikimedia, but that is a reverse proxy setup, not a forward proxy, and
those two workloads are inherently different with the forward proxy
setup much heavier on I/O due to the higher rate of cache replacement.

I know that there is several mobile phone operators running Squid proxy
services integrated in their access offering which maybe could compare
to what you are doing, but they are all also very silent about the
fact...

Other things I can note is that the usage of Squid on SUN platforms seem
to have decreased noticeably in the last years, with Linux & FreeBSD on
commodity servers from HP, Dell etc being the main platforms, even for
bigger installations.

Regarding threads the current answer is that Squid is not yet a good
multi-core performer. But it's an important area and something the
project wants to work on getting right. But it's also an area which will
require some substantial work and requires someone willing to sponsor at
least in part to get going.

Regards
Henrik Nordström
Squid Developer, Stockholm Sweden
http://www.henriknordstrom.net/

mån 2010-02-15 klockan 11:35 +0000 skrev Oborn, Keith:
>
> Hi all -
>
> We used to be a heavy user of commercial forward proxy/cache products,
> but with the demise of the NetApp line we stopped that activity.
>
> However, I'm now looking for an initial steer on a good large-scale
> forward proxy setup (multi-gigabit rates, hundreds of thousands of
> users). Unfortunately, our Polygraph rig was also scrapped some time ago
> (boxes died-), and it will take time and resource to build a new one.
>
> Sadly, the isp-caching and web-polygraph lists seem to be dead
> nowadays-.
>
> I'd be very interested in any numbers at all - most particularly on
> recent Sun kit, as it looks as if ZFS is a good bet for Squid. I must
> admit I alway hankered after testing a proxy/cache on a Thumper (X4540)
> because of the huge spindle density.
>
> At present, any numbers that get me within like a factor of two of
> actual performance on any modern X86 server hardware would be great (and
> perhaps any idea if using Sun T-series helps - does Squid like lots of
> threads?) - that will enable us to decide whether it is worth setting up
> to run our own detailed tests. I will, of course, post any test results
> we produce if we go down that road.
>
>
>
> ------------------------------------------------------------------------------
> Save Paper - Do you really need to print this e-mail?
>
> Visit www.virginmedia.com for more information, and more fun.
>
> This email and any attachments are or may be confidential and legally privileged
> and are sent solely for the attention of the addressee(s). If you have received this
> email in error, please delete it from your system: its use, disclosure or copying is
> unauthorised. Statements and opinions expressed in this email may not represent
> those of Virgin Media. Any representations or commitments in this email are
> subject to contract.
>
> Registered office: 160 Great Portland Street, London W1W 5QA.
> Registered in England and Wales with number 2591237
> ==============================================================================
>
Received on Wed Feb 17 2010 - 23:39:38 MST