Hi,
We're setting up Squid as a web cache for our 40+ employees.
As we recently desided to use multiple (unbundled) ADSL lines instead of a
$7000/month 2Mbps leased line, we would like to spread the load over the
Ethernet interfaces to which the ADSL modems are connected.
One way to do that, I guess, would be to have N+1 Squid instantiations
where 1 is the main cache and the other N are 'fake' caches that each have
a specific port for outgoing requests, thus making it possible (using e.g.
IPchains) to force that traffic over a specific interface.
My questions:
- is this the right approach for our sistuation of multiple lines to the
Internet
- does anyone here have any experience with such a set-up using Squid
- is this something that could / should be supported in Squid itself?
- does anyone has some tips on this setup?
As for the last remark, I obviously would love to say e.g.
httpd_accel_port 8080,8081,8082
where Squid would then randomly choose a port for the outgoing request.
- Stan
- Stan
Received on Fri Aug 04 2000 - 00:15:25 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:54:43 MST