Re: Very fast IP, no ATM...
-
To: (Christian Huitema)
-
From: Petri Helenius <>
-
Date: Thu, 5 Dec 1996 11:39:15 +0200 (EET)
-
Cc: "Paul Christ" <>, Peter Lothberg <>,
Christian Huitema writes:
> Paul,
>
> You asked me to repeat that:
> if not the telephone system itself- would Christian prepare
> an analysis saying "a Internet of `appropriate`
> capacity with RSVP and IPng would allow for concurrent
> AUDIO (as Jon said) of phone quality for everybody ..."
> The only problem I have with this phrase is the mention of RSVP. I don't
> think there is actually any major problem for providing a best effort
> service dimensioned to fit the needs of 600Mio internauts (or 6 billions,
> for that matter). I am convinced that some policing will be needed to
> make sure that usage is proportionate to investments (and subscription
> rates). I am however definitely not convinced that RSVP is what we need
> there. RSVP is about giving special priviledges to a few flows; we need
> to provision sufficient resources so that the average flow has adequate
> capacity, which is a different problem.
And for audio applications, redundancy can be used in the encoding for
the end result of packet loss tolerance without detectable quality
degration. This does not account for variable delay though.
Pete
|