NetNews Usenet Archive 1992 #19/NN_1992_19.iso/spool/comp/dcom/cellrel/395

Text File | 1992-08-27 | 4.0 KB | 79 lines

Newsgroups: comp.dcom.cell-relay
Path: sparky!uunet!pipex!warwick!str-ccsun!ccsun.strath.ac.uk!craa85
From: [e-mail] ( D.W.Stevenson)
Subject: Re: Future of IP routers
Message-ID: <[e-mail]>
Sender: [e-mail] (News account )
Nntp-Posting-Host: coll
Organization: University of Strathclyde
References: <[e-mail]> <[e-mail]>
Date: Thu, 27 Aug 1992 16:37:15 GMT
Lines: 66

In article <[e-mail]>, [e-mail] (Allen Robel) writes:
|> In article <[e-mail]>
|> [e-mail] ( D.W.Stevenson) writes:
|> > The addressing issues for ATM still haven't been resolved yet
|> > but I imagine it will be possible to address an individual service and
|> hence
|> > enforce policy restrictions.
|> >
|>
|> OK, I can see how this would work.  But, how do you (or others)
|> see segmenting the address space to reflect various service
|> offerings (e.g. 900 calls, 800 calls in todays phone system).

To go back to the initial subject of this debate, it was about whether IP
would be superceded by the ATM AAL layer as a network level protocol i.e.
will IP routers be needed in the future once there is a worldwide ATM network.
I think it will be a long time (if ever) before AAL protocols are used by hosts
as network level protocols, for the reasons discussed earlier. My earlier
postings were to point out that it's more efficient to run without the overhead
of the IP protocol but there is bound to be considerations other than efficiency.
(How do you define efficiency anyway?).

At the moment, all the effort is concentrating on using the AAL level as a
physical/link layer protocol and this inertia will probably remain. This keeps
the telcos happy because they are able to provide a carrier without knowing
what's going over it and it keeps the Internet people happy because the whole
idea behind the Internet is to allow traffic to pass over a variety of different
physical media, transparently (This raises another problem of how do you achieve
interoperability between IP based equipment and AAL only equipment if you do
away with IP routers)

I don't think anybody has got to the stage of trying to figure out all the
possible implications of a worldwide, seamless voice/video/data network, which is
what we are really talking about. It must be possible because it largely has been
achieved for a voice network.

ATM is still in it's very early stages and there
is still an awful lot of work to be done before it is anywhere near standardised.
It is quite possible that the telecomms standards bodies have looked at global
multimedia addressing, I don't know. Whether the telecomms addressing standards
will cover dividing data services down to the granularity that us computer people
require is debateable - this may be another reason for retaining IP (or it's
equivalent) as a network level protocol, since it keeps the division of
responsibility and culture between the carrier and the Internet people cleanly
cut.

I should point out that I am not actively involved in ATM work at all, so these
are my opinions formed from information gained via publically available sources,
and should not be taken as in any way authoritative - I know just as much (or
little) as anybody else who isn't actively involved in ATM work - perhaps
someone who knows about such things would like to comment.


|> everything else."  Much more useful would be the ability to
|> wildcard parts of the address that represented specific service
|> offerings (i.e. 1???1112222 might mean all interactive video services).

Assuming that a sensible policy on addressing is adopted, this functionality
can be easily built into a switch and probably in a more user friendly manner
than with the current bridge/router offerings.

--
Dave Stevenson                 [e-mail]
Computer Centre Communications        Tel : [phone] x3461
University of Strathclyde
Glasgow, U.K.