Re: No "TOP" of the docuverse [Was: URC usage scenarios ]

Michael Mealling (Michael.Mealling@oit.gatech.edu)
Wed, 28 Sep 1994 20:17:11 -0400 (EDT)

From: Michael.Mealling@oit.gatech.edu (Michael Mealling)
Message-Id: <199409290017.UAA00885@oit.gatech.edu>
Subject: Re: No "TOP" of the docuverse [Was: URC usage scenarios ]
To: connolly@hal.com (Daniel W. Connolly)
Date: Wed, 28 Sep 1994 20:17:11 -0400 (EDT)
In-Reply-To: <9409282316.AA04644@ulua.hal.com> from "Daniel W. Connolly" at Sep 28, 94 06:16:30 pm

Daniel W. Connolly said this:
> In message <199409282045.QAA14460@oit.gatech.edu>, Michael Mealling writes:
> >
> >The method I like is that what we do is each resolution server has
> >a unique id (what it is doesn't matter). We then have a directory of
> >services server that holds all the description URCs for each resolution
> >server.
>
> I suggest that any scenario involving a "root" or "top" server is
> doomed to fail.
>
> There are two key features to the URN/URC service: high availability
> and authentication. If you rely on a "root" server, you create a
> single point of failure, which conflicts badly with the goal of
> high availability.

I guess I should have elaborated more. There is no 'root'. Its simply
a service that any server can do if it wants to. Generally it should
be a server that is well connected and reliable. What I'm talking
about is the Directory of Servers service in whois++. Check out
draf-ietf-wnils-mesh-00.txt.

-MM

-- 
------------------------------------------------------------------------------
<HR><A HREF="http://www.gatech.edu/michael.html">
<ADDRESS>Michael Mealling</ADDRESS>
<ADDRESS>michael.mealling@oit.gatech.edu</ADDRESS></A>