The IP over ATM Mailing List Archive by date

[Craig Partridge <craig@aland.bbn.com>]

    On Fri, 29 Mar 96 03:17:25 EST Paul Ferguson wrote:

    >  Yes, ATM should change, in so many words.
    >
    >  IP, like it or not, is entrenched in the global environment [Internet].
    >  This is not likely to change.

    Is it likely that ATM will change?  This is an interesting question.  A lot
    of people seem to feel that ATM should change in order to better handle
    IP traffic.  While I don't think this is bad, I have to ask if this is
    realistic.  Is IP traffic likely to represent a significant percentage of 
    traffic carried on ATM networks (especially global networks)?  Or will IP 
    traffic be one of many, as Masataka pointed out.

Masataka was being ironic (note his smiley).

I think the trajectory (currently) is that IP will be the dominant user of ATM.
(A good sanity check would be to learn how much of today's Frame Relay traffic
is IP, since ATM is targetted to replace Frame Relay at some point in the
future).

So far as I can tell, only a small fraction of voice traffic will likely use
ATM, and in any case, voice is a shrinking proportion of the telecom network
(we're not quite at the point at which IP is the dominant user of long-haul
phone lines but some back of envelope calculations suggest it won't be too
many more years -- which is rather staggering).

As you note, the monkey wrench in this works is Video on Demand.  If (a) as
seems likely, the cable modems use ATM to deliver Video on Demand (that's
802.13's latest decision) and (b) people choose to use raw ATM rather than
send video over IP over ATM [a wide open issue right now] and (c) Video on
Demand takes off [another controversial topic in the cable industry right
now], then VoD is likely to have a large share of the traffic, maybe the
majority.  (Consider that one VoD channel will be about 6-8 Mbps sustained).

Craig