From: ccoprmm@oit.gatech.edu (Michael Mealling)
Message-Id: <199404281353.AA27294@oit.gatech.edu>
Subject: Re: Seperating URC format and URN->URC resolution
To: mitra@pandora.sf.ca.us (Mitra)
Date: Thu, 28 Apr 1994 09:53:31 -0400 (EDT)
In-Reply-To: <Pine.3.89.9404271553.B8856-0100000@pandora.sf.ca.us> from "Mitra" at Apr 27, 94 03:56:01 pm
Mitra said this:
> On Wed, 27 Apr 1994 rdaniel@acl.lanl.gov wrote:
> > Mitra's scenario currently locates the URI server(s) for a particular
> > publisher and gives us their IP address. However, this is not quite
> > enough. We have to know how to talk with the server we find at that
> > address. In order to support a range of alternatives in the experimentation
> > phase, how about we standardize on a well-known (to us :-) port and a
> > trivial service on that port which gives the client a list of ports and
> > services that can really do the URN->URL mapping. Browsers can look at the
> > list to see if they know how to speak the supported protocol(s) on that
> > URI server. Later on when we have standardized on a protocol, we can bump
> > up the version number of the trivial service and have it do the real thing.
>
> The big problem with this is speed, setting up a connection to a service
> is a non-trivial exercise, with around a 1 second penalty (which is why
> Prospero uses UDP).
While I havn't actually given this much thought. What about a
version of whois++ specifically restricted to run over UDP?
ie: no hold constraint, no interactive stream session. Just a UDP packet
with the search command + constraints from the client and a packet(s)
from the server with the result(s)?
It's just a though from someone who hasn't had a chance to work with
UDP all that much...
-MM
-- ------------------------------------------------------------------------------ Michael Mealling ! Hypermedia WWW, WAIS, and gopher will be Georgia Institute of Technology ! here soon via MIME. Your view of the Michael.Mealling@oit.gatech.edu ! internet is about to change completely!