RE: multiple squid with load balance

There are a number of ways to do this, with varying levels of availability
and load balancing or sharing.

1) DNS round-robin - this will share the load across multiple addresses, but
has no way to compensate for errors, so if one out of two servers is down,
50% of your users' requests will fail.
2) Auto-config script - you can set up JavaScript that you're users will
download and set their proxy preferences based on. This can say "Use proxy1
unless its down, and then use proxy2". Its more of an availability thing
than a load-balancing thing.
3) L4 switch-type - this is a device such as a Cisco LocalDirector or Alteon
ACEDirector which sits between the main network and a farm of servers and
redirects traffic across a group of similar services. This can provide both
true load-balancing and high availability.

I've just recently finished a paper on this for a class I was taking, and
for availability and load distribution, I found that the third type had the
most benefits. Furthermore, it makes scaling much easier, since you can
simply add another box to the farm when you outgrow your current setup. The
biggest drawback to it, though, is purchase price, since L4 switches aren't
cheap, but if you're dealing with a large environment with high demand and
little tolerance for down-time, it is the best way to go.

Jon

> -----Original Message-----
> From: Ricardo Clementino [SMTP:ricardoc@folhasp.com.br]
> Sent: Thursday, July 08, 1999 3:17 PM
> To: squid-users@ircache.net
> Subject: multiple squid with load balance
>
> Hi everybody,
>
> Anybody can tell me what i need to do to configure N squids with proxy
> that my clients can access using only one IP address
Received on Fri Jul 09 1999 - 07:26:18 MDT