[e2e] Cannara's views
vjs at calcite.rhyolite.com
Fri Apr 13 10:16:54 PDT 2001
> From: "David P. Reed" <dpreed at reed.com>
> X.25 networks or the ATM SVC networks. And if congestion turns out to be
> better controlled by deploying twice as much fiber as needed, that seems to
> me to be both theoretically and practically sound.
I think there's something else at work here. Part of the unstated
world view of the Internet seems to me to be decentralization. With
a central control board in charge of everything, you don't need (in
theory) to worry about interoperability and heterogenity; you just
outlaw them. On the other hand, if you don't believe in central
benevolent dictators, you start by assuming heterogeneity and that
forces you to worry about interoperability.
The complaints over the years from the other gentleman sound much like
the complaints about CSMA/CD. Recall the endless claims that CSMA/CD
collisions waste vast quantities of bandwidth from people who evidently
never quantified bits spent in collisions even at 100% collision rates.
The same people never tired of claiming that 802.5, FDDI, and 100VG-AnyLAN
had better latency guarantees and utilization than CSMA/CD because
both were explicitly controlled. They evidently also never put numbers
to the incredible, absolutely useless durations of the latency guarantees
of 802.5 and FDDI (6 seconds!) and the terrible utilization of
100VG-AnyLAN because of the continual mother-may-I exchanges with the
center of the net.
Some people are deeply offended by democratic or free market systems.
If you prefer, they demand that every mechanism be controlled by
explicit mechanisms, fields or packets. The IETF is not immune to such
things as the PPP BACP nonsense-protocol demonstrates. (It's nonsense
because it involves exchanging packets carrying information that is
available to both sides simply by counting bits, and commands that would
be ignored if they were actually sent.)
Vernon Schryver vjs at rhyolite.com
More information about the end2end-interest