Re: thoughts on memory usage...

From: Martin Hamilton <martin@dont-contact.us>
Date: Mon, 18 Aug 1997 11:22:53 +0100

-----BEGIN PGP SIGNED MESSAGE-----

David Luyer writes:

| An efficent string-matching algorithm is essential for this kind of
| compression of course, but it should be achievable without too much rise
| in CPU usage. CPU power is much cheaper than I/O or memory bandwidth most
| of the time anyway, and various systems have relatively low limits on
| physical memory configurations. Also, if one was sufficiently optomistic,
| squid caches could communicate their version number and, if compatible,
| communicate with compressed URL strings.

Just to note that (for us at least :-) the CPU usage by Squid isn't a
problem. For instance, on a 200MHz Pentium Pro box servicing
~1,000,000 combined HTTP and ICP requests per day, we're seeing idle
times in the region of 94%. A little bit more CPU usage might not
hurt, especially if it reduced memory consumption...

Cheerio,

Martin

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3i
Charset: noconv

iQCVAwUBM/gietZdpXZXTSjhAQEp8wP/QHD8JzVCobMqrAlXM64HveoAB5mkuNIJ
Ds8Pw7Qvp2fg41zm0fY0MDVpLxmnzM4CZdBU2KC5ESMxfOgISh62+zGp6Uuff5dm
vHQUi2Lm0V3s1t/cTnb/jNmHvwo/pAex3U8Lrl1WkGWieTf31EIP9Aal48vUPdfR
ADb5AMIXUzo=
=QuOd
-----END PGP SIGNATURE-----
Received on Tue Jul 29 2003 - 13:15:42 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:11:22 MST