Re: AW: [squid-users] cache performance

From: Squid Support (Henrik Nordstrom) <hno@dont-contact.us>
Date: Tue, 7 May 2002 12:41:37 +0200

Ulrich Walcher wrote:
> You could also take a look on the pages mentioned in the squid.conf:
>
> # For more information about the GDSF and LFUDA cache replacement
> # policies see
> http://www.hpl.hp.com/techreports/1999/HPL-1999-69.html # and
> http://fog.hpl.external.hp.com/techreports/98/HPL-98-173.html.

If reading these papers, be sure to take note of how the policies behaves
when applied to sufficiently large caches.. (all behave approximately the
same).

Looking at different policies mostly makes sense for very small caches, such
as main memory caches. However, Squid's implementation of cache_mem has much
larger problems making the memory policy question almost irrelevant.

-- 
Basic free Squid support provided thanks to MARA Systems AB
Your source of advanced reverse proxy solutions or customized
Squid solutions. http://www.marasystems.com/products/
Received on Tue May 07 2002 - 04:42:01 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:07:56 MST