Re: [squid-users] relation between http req/sec and no of simulatneous accesses

----- Original Message -----
From: "Steve Snyder" <swsnyder@home.com>
To: "khiz code" <khizcode@yahoo.com>; <squid-users@squid-cache.org>
Sent: Friday, October 12, 2001 3:30 PM
Subject: Re: [squid-users] relation between http req/sec and no of
simulatneous accesses

> On Friday 12 October 2001 12:16 am, khiz code wrote:
> > Hi all
> > ive been stress testing my squid boxes using ummm .. wget (mutilple
> > copies) and web stress tools
> > my cache could go upto abt 70 req/sec
> > what i would like to understand is that 70 req/sec corresponds to
abt
> > how many simultaneous client connections ..i.e how many clients can
the
> > cache serve at any point of time
>
> I don't have a specific answer to your question, but I can offer that
I've
> seen plenty of pages with 30+ objects in them. If only 2 users
requested
> pages like this simultaneously, your 70 request are just about
consumed.

Browsers will use a maximum of 2 http connections to a single cache at
once (for http/1.1 browsers that is, http/1.0 are allowed more, but most
1.0 browsers only use 4 connections.

So when a use hits a page with 30 objects, they get (assuming the
objects are equal in size) 15 on one connection and 15 on another. The
entire object transfer rate would need to be 1/15th of a second, or
approx 70ms for the requests per second count to be pushed up to 70/seq
by two users.

If you look at your stats page, I'm sure you'll see that misses take a
little more than 70ms to satisfy :].

Rob
Received on Fri Oct 12 2001 - 00:43:27 MDT