You have several options:
You can use the Squid builtins to share caching information and/or
configure them into a parent/sibling relationship.
You can spread load several ways using proxy autoconfiguration files.
http://www.squid-cache.org/Doc/FAQ/FAQ-5.html
http://home.netscape.com/eng/mozilla/2.0/relnotes/demo/proxy-live.html
You can use a DNS round-robin structure. It's not very accurate
load balancing, but it's easy. In a DRP scenario when one
of your caches fail and you yank it out of the pack, propagation of
that information can be slow and meaningless because of ARP cache
and the particulars of your environment.
You can pack a bunch of proxies behind a Linux Virtual Server (LVS)
http://www.LinuxVirtualServer.org/. I think it looks schweet. It
will do redendancy and load-balancing.
You can buy a Cisco Local Director, which is basically the same
functionality as the LVS but very likely faster and more scalable.
http://www.cisco.com/warp/public/cc/cisco/mkt/scale/locald/index.shtml
For those of you who are celebrating a holiday soon, please remember
that alcohol and root don't mix. Have the forethought to have a
dedicated sober Squid admin. For your safety and ours.....
Received on Tue Dec 21 1999 - 09:14:47 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:50:04 MST