Re: Squid memory footprint (was: MemPools rewrite)

From: Alex Rousskov <rousskov@dont-contact.us>
Date: Mon, 6 Nov 2000 15:30:59 -0700 (MST)

On Mon, 6 Nov 2000, Henrik Nordstrom wrote:

> Incremental digest does not really make sense here as the digest are
> never exchanged. I was more thinking of a gradually decaying digest.
>
> Objects gets added to the digest when they enter the cache, and deleted
> from the digest by clearing some bits when they are evicted from the
> cache. To make sure bits never get stuck in the digest there is a
> function which makes sure that no bit can stay on for ever. A simple
> slow scan clearing bits should be sufficient for this.

That's what I meant by "incremental digest" , i.e., digest that is
updated incrementally rather than once. Pei Cao's Cache Summaries are
this kind of digests -- they allocate a few bits per digest "entry" as
a counter of additions/deletions...

Alex.
Received on Mon Nov 06 2000 - 15:31:07 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:12:56 MST