Date: Mon, 7 Nov 1994 22:48:13 +0100
Message-Id: <199411072146.AA254294786@relay.acns.nwu.edu>
From: Albert-Lunde@nwu.edu (Albert Lunde)
To: Multiple recipients of list <www-proxy@www0.cern.ch>
Subject: Re: Chaining proxies
At 8:18 PM 11/7/94 +0100, N.G.Smith wrote:
>Of course this is a configuration nightmare, and some fallbacks need to
>be built in to handle error situations, (just revert to Direct), but
>that is just icing on the cake. With a good implementation this could
>be the basis for URNs. (In fact URNs could just be URLs with the host
>part signifying the publishing site rather than the site to fetch
>information from).
[...]
>>@ I believe that this is the way to go. There is a definite, but
>>@ technically simple, step going from the current CERN server to one that
>>@ supports dynamic choice of secondary proxy/cache. Once this is in
>>@ place, providing a URN type service should be significantly simpler
>>@ than making the whole URL->URN transition in one go.
[...]
>As I say above, with good caching strategies, URLs are URNs. It's up to
>the caches between the user and the publisher to decide how they resolve
>the situation.
>From what Mitra said at the WWW conference, the URN "testbed"
implementation, that is being used by the URI working group developing the
standards, will use a modified proxy server as the immediate agent to
resolve URNs to URLs.
The resolution server the proxy server uses to ultimately resolve URNs will
be located via DNS, (i.e. uri.path.net will be the resolotion server for
path.net)
For more about this, see Mitra's web page:
<http://www.path.net/mitra/>
None if this is final, and it may not all apply to all implementations.
---
Albert Lunde Albert-Lunde@nwu.edu