[1009] in Commercialization & Privatization of the Internet
Re: some dumb questions from the gallery
daemon@ATHENA.MIT.EDU (Edward Vielmetti)
Mon Jul 15 22:05:29 1991
To: tal@Warren.MENTORG.COM
Cc: nren-discuss@psi.com, com-priv@psi.com
In-Reply-To: Your message of Mon, 15 Jul 91 21:02:53 -0400.
Date: Mon, 15 Jul 91 22:02:26 EDT
From: Edward Vielmetti <emv@ox.com>
What if email becomes as pervasive as the telephone? Will we have
enough bandwidth to carry ?
Well, by my reckoning, e-mail is easier to carry than voice; it's
store and forward, after all, and it's generally more compact since
you can compress things and you aren't constrained by real time.
If the current PSTN is adequate for voice, then it's probably going
to be adequate (along with metro nets and bypass fiber) to handle all
of the e-mail you and I and Aunt Gail can type with our own damn fingers.
The problem is that the net is being asked to carry both the moral equivalent
of 18-wheel monster trucks (expensive big-science real-time video or
massive supercomputer graphic visualization) along with the ordinary
passenger car traffic (e-mail, netnews) and it's hard to engineer a network
that's adequate to both tasks that very many people can afford.
I was reading some stuff about highway engineering; seems that the
durability and capacity of a highway is determined by the thickness and
quality of construction of the pavement as well as the number of lanes and
the width of those lanes. Big trucks put a disproportionate load on the
highway (akin to the 3d power of per-axle weight), and the current tax
structure doesn't provide the right incentives to minimize axle weights;
similarly, the big-science applications for real time packet video or
world-wide distributed file systems can put an enormous strain on networks
and systems which would cope just fine for e-mail with a national
infrastructure built out of something like SMDS or Frame Relay.
--Ed