Yep. So far the only realistic way to move massive amounts of
long-distance traffic is fiber.

>And then there is all that PSTN infrastructure that will
>have to be replaced...

Yep. It's only money. They already have rights of way.

Ah... but do they have -all- the rights of way & fiber?
Are there alternatives to fiber? Those are the 64ruble questions...

If they do, and no alternatives, then a terabit router is worthless.
I can't get the infrastructure to support it. If the PSTN is the
only game in town, then its ATM or nothing.. and ATM won't cut it
at those rates.


bmanning@ISI.EDU writes:

If the PSTN is the only game in town

The PSTN is dead; the only thing is that the body is still
giving off lots of heat so fibre owners haven't begun
giving up on it yet. Some of them even appear to be
betting that in three days it'll get up and start walking
around again.

Real money is in computer mediated and computer-to-
computer communications, and it's that money which is
sustaining the last vestiges of metabolism in the PSTN.

then its ATM or nothing.. and ATM won't cut it at those

Well it's not surprising that telcos still hope to save
the PSTN by selling only ATM like in silly cases where a
clear-channel colocation-facility-to-colocation-facility
SONET run is ignored as a sales option while an ATM PVC of
the same general bandwidth takes a thousand-mile detour to
and from the nearest ATM switch...

Moreover, I'm told that currently on offer from one telco
is a OC3 SONET pipe for only ten times the price of a DS3.
However, if you want a PVC following the exact same path,
the circuit is only a little more than double DS3.

(Ironically you can currently do 4:3 SONET/SDH inverse
muxing giving you an OC3/STM-1 from 4xDS3 (plan your DS3s
along divers paths, kids!). Of course, you can just use
various router vendors' strategies for load-balancing
among DS3s and use all 4 DS3s as they are, for that matter...)

Finally, to tie in a reply to your reply to my previous
message, finding ways to light up fibre is not so
difficult; at worst it all comes down to massive amounts
of MUXing. While I happen to prefer the approach of a
combined SONET MUX/IP router to stick at each end of a
pair of soliton-containing fibre strands, Vadim Antonov's
approach is also tractable and attractive, and has some
interesting benefits that could be seen as side effects
wbut which he has turned into a real additional-value
feature within his proposed product.

In any event, let me just say that I disagree with you
that "stringing is easy" -- replacing all the crap in the
ground in the U.S. now with fibre that has decent
chromatic dispersion properties and as little dependence
on electro-optical conversion as possible is going to be a
long, expensive and labour-intensive process.

Hence the window of opportunity for Hans Werner von Braun
and his amazing rocketships of death, and similar
techologies I'd rather not have to depend on. :slight_smile: