Date: 29 Sep 94 08:13:52+0100
From: pays@faugeres.inria.fr
To: Michael.Mealling@oit.gatech.edu, francis@cactus.slab.ntt.jp,
Subject: Re: URCs and URL resolution
Message-Id: <780822832.16771.0-faugeres.inria.fr*@MHS>
Concerning the scalability of whois++ (in fact the
suitability scalability of centroids):
1. As indicated by PF, one crucial thing is the attribute choice,
wrt to the type of queries that will be issued. It is not
really a technolgy matter but more an organisational
and operational matter
2. Right! there are several question marks about centroids on a large
scale, choice of he centroid mesh aso...
As, on the other hand, if we have no proof yet concerning centroids,
we (people with experience in directory services) sure have
proofs that existing technologies (such as X.500) will NOT
scale, AND have many hope that centroid can do it.
THUS, many of us are about the launch a large scale centroid
worldwide pilot, in order to have facts and experience about
centroids rather thab belief and disbelief.
a/ the pilot is a centroid pilot (not a whois++ pilot) as it
will include at least whois++ and SOLO
b/ the 2 major apps that will be studied will be white pages
AND URN URC URL name service
c/ I will open the mailing list in the next few hours,
those interested in an active participation are welcome.
just drop me a line
d/ I will try to set-up a preparatory face to face meeting
this autumn in Paris (date to be decided) 2 to 3 days long.
NOTA: we are not yet another standard making group. The
purpose of the activity IS experimentation. As the key people
of Whois++ and Solo will be involved they may later evolve
their specs in the light of the pilot results, but this is not
included in THIS pilot activity.
Participation seeked is an ACTIVE participation: deployement
and operation of a centroid mesh for realistic applications
(though certainly not an operational service)
-- PAP