[squid-users] how to achieve squid to handle 2000 concurrent connections?

[Amos Jeffries]

On 28/04/2015 6:41 p.m., Abdelouahed Haitoute wrote:
> Hello Amos,
> 
> Thank you for your reply.
> 
>> Consider removing the "cache deny all" when you get this into
>> production.
> 
> 
> We’re expecting all requested files to be unique. So is it still wise to cache these requested files? Are there other advantages of enabling caching?
> 

For truely unique fetches it will not have much use. Disk cache would be
a negative due to latency. RAM cache would be not adding benefit, but a
small (few MB) amount would not be reducing performance either.

For other benefits, you get caching for any unplanned non-unique
requests. eg. DoS events or changes in the applicaton design.

Its also an early indicator of design problems. If your unique files
cannot cope with a cache existing in their pathway something is terribly
wrong with the headers being output by the server.

Amos