Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!know!zaphod.mps.ohio-state.edu!wuarchive!mailrus!accuvax.nwu.edu!nucsrl!telecom-request
From: tjo@its.bt.co.uk (Tim Oldham)
Newsgroups: comp.dcom.telecom
Subject: Re: Data vs Voice
Message-ID: <12624@accuvax.nwu.edu>
Date: 25 Sep 90 16:35:14 GMT
Sender: news@accuvax.nwu.edu
Organization: BT Applied Systems, Birmingham, UK
Lines: 45
Approved: Telecom@eecs.nwu.edu
X-Submissions-To: telecom@eecs.nwu.edu
X-Administrivia-To: telecom-request@eecs.nwu.edu
X-Telecom-Digest: Volume 10, Issue 680, Message 8 of 10


In article <12542@accuvax.nwu.edu> adiron!tro@uunet.uu.net (Tom Olin)
writes:

>What happens if all those callers simultaneously break into song or in
>some other way push their duty cycles up to 100%?  Do they start
>losing parts of their conversations?  Or do pieces merely get delayed?
>Or does something else happen?

As another contributor has said, the technique of zero compression
isn't prevalent on land nets. It is, I believe, prevalent on satellite
links; certainly sometimes I get completely dead patches on a noisy
phone line, which I have always contributed to zero compression.

So, am I correct in this assumption? If I'm right, what happens when
satellite virtual-circuits *do* start using more bandwidth. For
example,

Take a satellite which has 64Kbps bandwidth (it's a pretty crap
satellite :-):

time t: I'm on the line to the US from the UK, and it I'm on a
        zero-compressed satellite link. I'm put on hold, and keep my
	mouth shut. No bandwidth being used.

time t+1: Another person decides to call the US. Whatever-it-is decides
	that statistically (assuming a 50% zero compression winnitude),
	and currently actually, there is sufficient bandwidth to route
	the call via the satellite. It does so, and the 2 parties are
	connected. They start talking. No problem; plenty of bandwidth,
	as I'm still not saying anything.

time t+2: All four parties start screaming blue murder at each other. We
	need 128Kbps bandwidth. Which the satellite can't deliver.

What happens? Is it just a case of the designers allowing for
statistically very low data loss, given that the numbers are very much
bigger for a typical satellite? If not, what techniques are used?

Yeah, I work for a telecomms company, but I'm not directly involved in
telecomms. Just interested.


Tim Oldham, BT Applied Systems. tjo@its.bt.co.uk or ...uunet!ukc!its!tjo
Living in interesting times.