[11123] in Commercialization & Privatization of the Internet
Re: uniform policies at the NAPs
daemon@ATHENA.MIT.EDU (Hans-Werner Braun)
Tue Mar 22 06:00:53 1994
From: hwb@upeksa.sdsc.edu (Hans-Werner Braun)
To: swb1@cornell.edu (Scott W Brim)
Date: Sun, 20 Mar 94 9:16:10 PST
Cc: cook@path.net, com-priv@psi.com
In-Reply-To: <199403190440.XAA13662@postoffice3.mail.cornell.edu>; from "Scott W Brim" at Mar 18, 94 11:39 pm
Scott:
>Gordon, like, for sure they don't have to be uniform as long as
>everything *works*. If the customers are happy nothing else matters.
>I would call this the responsibility of the routing arbiter.
I would even go further than that. With the design of the NAPs,
including a level 2 substrate, non-homogeneous policies are explicitly
enabled, with the RA helping out as they can and as need be. May be two
NSPs want full backup of each other via the NAPs (y'know, some just
want robustness and make their customers happy), two other NSPs may not
ever want to see a packet from each other (turf protecting or
whatever). Others, for that matter, may want to avoid NAPs at all cost
to begin with, be it to send a political signal, to avoid NSF supported
facilities as they may believe for example that NSF regional networks
make it harder for them to argue why some end customer like a
university should buy from them, or whatever.
This sure won't all go smooth over the next few years, given growth,
diverse interests and policies, lots of players, and just generally
noisy people. I suspect a number of OK Corral scenarious. We have seen
many of them already, and we will see many more of them as in the
future the arguments won't just be by purity of heart, religious
fanatism, personal fights, or plain "I care about the future of the
whole Internet," but by plain and simple "I want to have a bigger share
of the market for my own gain." Haven't we, possibly between the lines,
be seeing more and more of the latter lately? That's probably just life
and normal.
Personally, I am delighted to still see at times people who care about
the future of the Internet network at large and as a system. Then
again, sometimes, including watching some discussions, I wonder whether
they get fewer and farer in between. Sigh. Hey, Scott, may be we need a
new sandbox and collaborate on that? Like we did when the NSFNET was
still young and the tender age of 56kbps? You do all the dirty work
again and I have all the fun?
Hans-Werner
