Re: Squid memory footprint (was: MemPools rewrite)

From: Andres Kroonmaa <andre@dont-contact.us>
Date: Mon, 6 Nov 2000 14:19:01 +0200

On 6 Nov 2000, at 6:03, Joe Cooper <joe@swelltech.com> wrote:

> > Imagine a setup: 10 caches, loadbalanced with L4 switch, seen as a
> > single cache by clients. Each box takes 1/10th of the http load, but
>
> I'm confused. Are there any L4 switches that don't provide a means to
> schedule cache usage based on a hashed value or some other network
> division to avoid redundant data? Even LVS now has a hash based
> scheduling algorithm so that there is never any redundant data on the
> balanced caches.
>
> ICP shouldn't come into the equation at all in an L4 balanced network,
> unless there are L4 switches that can't intelligently select a cache?

 You are talking about transparent caching. If you have nontransparent
 caching, then all those hashes can be thought of only inside proxy, or
 based on URL, but then you'd need a L7 switch. You'd need to intercept
 TCP session, wait and gather URL information (multi-packet, btw), and
 then make decision on which cache to pick. This cannot be done fast in
 hardware, so its not a job for a switch.

 Besides, destination-based loadbalancing can easily overload a single
 cache, so its not very optimal solution.
 L4 balancing also implies dead-cache detection and redirection to other
 caches. So, eventually destination IP based loadbalancing is not at all
 attractive.

> ICP has some good uses, but use when balancing caches is unwise, I
> think.

 ICP is all we have now. Digests seem better, but do not allow instant
 knowledge of peer contents.

------------------------------------
 Andres Kroonmaa <andre@online.ee>
 Delfi Online
 Tel: 6501 731, Fax: 6501 708
 Pärnu mnt. 158, Tallinn,
 11317 Estonia
Received on Mon Nov 06 2000 - 05:22:03 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:12:55 MST