lör 2006-04-29 klockan 12:10 +0100 skrev Michael Pye:
> This depends on a large number of factors, but there's no reason why you
> could not get several hundred requests per second, if not several
> thousand requests per second being handled.
Most reverse proxy setups I have seen in production peaks around 500 to
1500 requests/second of real life traffic depending on type of workload
on the site, how much the Squid code have been tuned and hardware.
Synthetic test workloads can get a lot higher with less work..
Note: Forward Internet proxying is generally much harder to handle than
reverse proxying as the diversity of the requests is much larger in
forward Internet proxies.
Regards
Henrik
This archive was generated by hypermail pre-2.1.9 : Mon May 01 2006 - 12:00:02 MDT