Re: V2.3 & Cache Digest

From: Jens-S. Voeckler <voeckler@dont-contact.us>
Date: Fri, 30 Jun 2000 15:59:21 +0200

"Jens-S. Voeckler" wrote:
>
> ]Use both constraints in conjunction, and add a good pinch of salt to
> ]make up for inaccuracies, e.g. more than 500 MB per day or more than
> ]3 mio requests per day.
>
> Make that 1 .. 2 mio requests per day.

And a final thought: The request constraint does not yield very good
numbers - at least for larger meshes. 2 mio requests per day would
imply a mean of 23 ICP per second and client. Unless you want to die
in an ICP storm, or you have only few clients, digests are a way out
of that dilemma, as is limiting the members in ICP destination groups.

Also, getting request number from parent caches is difficult, if CDs
are already employed and the clients configured "no-query". The
clients may peer with other caches besides your parents and there may
exist other limiting configurations specifics. You do not see at the
parent, how often your digests was used within the client to look up
some object. Maybe that can be found in your clients mgr:* material?

-- 
Le deagh dhùrachd,
Dipl.-Ing. Jens-S. Vöckler (voeckler@rvs.uni-hannover.de)
Institute for Computer Networks and Distributed Systems
University of Hanover, Germany; +49 511 762 4726
Received on Fri Jun 30 2000 - 08:02:41 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:54:15 MST