Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!samsung!umich!mailrus!accuvax.nwu.edu!nucsrl!telecom-request
From: tjr@ihnet.att.com (Thomas J Roberts)
Newsgroups: comp.dcom.telecom
Subject: Re: Phone Calls at the Speed of Light
Message-ID: <3544@accuvax.nwu.edu>
Date: 5 Feb 90 15:42:54 GMT
Sender: news@accuvax.nwu.edu
Organization: AT&T Bell Laboratories
Lines: 43
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 77, message 6 of 5


From article <3396@accuvax.nwu.edu>, by boomer@athena.princeton.edu 
(Don Alvarez):

> In article <3335@accuvax.nwu.edu> eli@pws.bull.com writes:
>>  [table of signal velocity characteristics in wire deleted]
> Since the slowest speed listed here is about 94% C, and one can only
> assume wires have gotten better, not worse in the last 47 years, we
> clearly should all drop our fiber optic lines and go back to copper.
> (God, how I hate waiting for those 20ms delays!)
 
> -don

Of course, the delay in the wire is only a small part of the total
delay.  Every modem (digital communication only) adds >1000
microseconds of delay.  Every analog amplifier adds several to many
microseconds of delay.  Every analog Frequency Division Multiplexor
adds several to many microseconds delay.  Every digital regenerator
adds up to a microsecond of delay.  

Every Analog->Digital or Digital->Analog conversion adds up to 125
microseconds of delay. Every time-slot-interchange within a digital
switching system adds 125 microseconds of delay. Every satellite hop
adds >100,000 microseconds of delay (but most of us don't have to
worry about this). After you add all of these up, you then get to
double the delay if you're concerned about the round-trip delay (and
normally that's what is of interest).

Fiber is very different, as the regenerators add only a few to many
nanoseconds of delay, and they are spaced further apart. But the
signal propagation in the fiber is slower than that of a wire or of a
radio channel.

The analysis of round-trip delay on a telecommunications channel is
VERY complicated, and is extremely route dependent. If your path
traverses a packet network, things can get downright bizarre.  Note
that signal delay is only a minor consideration when telcos engineer a
transmission path (cost is the major factor, capacity, ease of
maintenance, and reliability overwhelm delay considerations).


Tom Roberts
att!ihlpl!tjrob