Re: [squid-users] squid source problem

From: Colin Campbell <sgcccdc@dont-contact.us>
Date: Thu, 23 Aug 2001 07:22:12 +1000 (EST)

Hi,

On Wed, 22 Aug 2001, fatemeh bishemar wrote:

> Thanks a lot for your beautifull answer, but how one
> process ( child
> process) can process these two much requests that
> arrives from client in a
> good speed, and how the requests don't made mistakes
> with each other?

Squid is a single threaded process that relies on the select() or poll()
library calls to handle multiple connections simultaneously. I believe all
I/O is non-blocking too. It's an alternative to forking a child to handle
every connection. It requires more programming to make it work properly
but you avoid the overhead of process creation as in the fork() model.

Colin
Received on Wed Aug 22 2001 - 15:22:18 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:01:53 MST