Re: [squid-users] Is Squid slow?

Joe Cooper wrote:
[...]
> http://www.measurement-factory.com/results/public/cacheoff/N03/auto/all/rep.rptm.mean.save.ratio.html
>
> Swell was 11th out of a field of about 30 entries and Duane's Squid box
> was an even more respectable fourth place. This field is generally an

I believe you're reading the graph incorrectly. The response time
improvement graph shows how much of a benefit the cache provides.
A 0% improvement means no benefit. So, on this particular graph,
the entries on the right side are performing better than the ones
on the left.

As a sanity check, take a look at the hit time graph:
http://www.measurement-factory.com/results/public/cacheoff/N03/auto/all/hit.rptm.mean.html
The two black bars are Squid, and don't fare particularly well.
The situation is similar for the miss times, at
http://www.measurement-factory.com/results/public/cacheoff/N03/auto/all/miss.rptm.mean.html

This is not a slam on Squid, but rather just a correction of an
interpretation of the graph. Again, disclaimer: I work for a
caching company.

-Vivek
Received on Tue Dec 04 2001 - 20:37:46 MST