W3C home > Mailing lists > Public > www-talk@w3.org > November to December 1996

Re: Mirror Negotiation

From: Jon Knight <jon@net.lut.ac.uk>
Date: Tue, 10 Dec 1996 19:08:31 +0000 (GMT)
To: Chanda Dharap <chanda@mowgli.PRPA.Philips.COM>
cc: Peter J Churchyard <pjc@tis.com>, www-talk@w3.org
Message-ID: <Pine.SUN.3.95.961210185305.12894T-100000@weeble.lut.ac.uk>
On Tue, 10 Dec 1996, Chanda Dharap wrote:
> Has anyone given thought to some sort of a hierarchy of caches ? Maybe
> topical caches as opposed to regional ? Caching could be a combination
> of server-side Proxy caching as well as caching at the hierarchy.
> 
> The hierarchy of caches would be neither at the server nor at the
> client, somewhere in between.

The UK already has a (sort of) cache hierarchy; many JANET sites have a
campus cache that peer with a few neighbours and use the HENSA cache as a
parent.  The idea is that if everybody in .ac.uk did this, the only
transatlantic traffic would be coming from the HENSA cache (which has its
own transatlantic link), leaving the rest of the Fat Pipe free for
un-cachable traffic, telnet, smtp, MBONE, etc, etc.  As many of the
academic sites are likely to be pulling similar sets of documents at
around the same time, there is a good chance that the cache hierarchy will
have a reasonably high hit rate.  We're certainly better off with it than
without it.

On a more local level, we also distribute our ROADS subject gateway
support software with an option to run the link checking module through a
cache in order to populate it on a subject basis; I don't know whether any
of the services that use the software have actually enabled that option
though. Others are working on attaching an indexing robot such as Harvest
or NWI to the ROADS database to pick up related resources and of course
these could also be pulled through the subject basd proxy cache to
pre-load it.  There's quite a lot of interest within the Access to
Network Resources strand of the UK Electronic Libraries programme about
the use of subject caching so we're hopeful that someone will make use of
these features eventually.

Tatty bye,

Jim'll

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Jon "Jim'll" Knight, Researcher, Sysop and General Dogsbody, Dept. Computer
Studies, Loughborough University of Technology, Leics., ENGLAND.  LE11 3TU.
* I've found I now dream in Perl.  More worryingly, I enjoy those dreams. *
Received on Tuesday, 10 December 1996 14:08:50 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 27 October 2010 18:14:20 GMT