Re: Q: Hit Rate calculation

At 10:17 AM 7/1/97 +0200, Jan Torreele wrote:
>Hi all,
>
>I have a question about the calculation of the hit rate of a cache. Using
>different tools (i.e. Cache Manager, Squidstats, Squidtimes, spa,
>manual log analysis, etc.) I obtain wildly varying values for the hit
>rate of our cache. Which of these should be considered 'correct'? What I
>would like to know is what percentage of requests (TCP) to my proxy is
>actually being served from the cache. How do other people caculate this?

This depends on what you call a hit. If you just say TCP_HIT is a hit and
TCP_IMS_HIT etc isn't (IMS does a request to the site and gets back use
local cache) the hit-rate is not to good. TCP_HIT is only given to fresh
entries. A stale entry can still be valid and result in a local cache reply
from the www-server but strictly speaking it accesses the web so strictly
speaking it isn't a hit.

My opinion is that IMS_HIT etc is still a hit because the data received from
"The Web" is limited compared to a MISS.

Squidstats counts only TCP_HIT and UDP_HIT and all the rest is a miss (I'm
running a modified version so IMS etc is also counted as a hit)

Marc
---------------------------------------------------------------------
Marc van Selm
NATO C3 Agency
Communication Systems Division, A-Branch
E-Mail: marc.van.selm@nc3a.nato.int
---------------------------------------------------------------------
Private: selm@cistron.nl, selm@het.net, http://www.cistron.nl/~selm
Received on Tue Jul 01 1997 - 04:19:17 MDT