W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > May to August 1996

Re: When to make objects uncacheable ?

From: Luigi Rizzo <luigi@labinfo.iet.unipi.it>
Date: Tue, 20 Aug 1996 20:16:52 +0200 (MET DST)
Message-Id: <199608201816.UAA18159@labinfo.iet.unipi.it>
To: Hull Chris <Chris.Hull@fmr.com>
Cc: tai@hplb.hpl.hp.com, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com, ircache@nlanr.net, l.vicisano@iet.unipi.it, Luigi Rizzo <luigi@labinfo.iet.unipi.it>
> Well, actually the current hit rate is zero.  Due to sizing
> problems we have had to turn off caching a few weeks ago.  That's
> why I found out about this list, and also why I wrote some code to
> analyze 600MB of log files.  I'm looking to implement a multi-level
> cache, distributed over three sites, but today I am just collecting
> data to model the sizing issues.

We have a techreport:

	http://www.iet.unipi.it/~luigi/caching.ps.gz

which discusses some issues in the relation between cache size and
hit rate.

Actually, we have also done some simulations on the performance of
hierarchical and/or cooperating caches; a report on that is not
available yet, but the bottom line is the following:

N cooperating caches (a la Harvest) with disk capacity D each work
almost as a single cache with N*D disk space (for N not too large,
let's say N < 10).  The communication between caches is a small
fraction of the overall traffic (because it only affects documents
which are accessed more than once).

	Luigi
====================================================================
Luigi Rizzo                     Dip. di Ingegneria dell'Informazione
email: luigi@iet.unipi.it       Universita' di Pisa
tel: +39-50-568533              via Diotisalvi 2, 56126 PISA (Italy)
fax: +39-50-568522              http://www.iet.unipi.it/~luigi/
====================================================================
Received on Tuesday, 20 August 1996 11:07:14 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:32:08 EDT