[1253] in Commercialization & Privatization of the Internet
Re: technical details
daemon@ATHENA.MIT.EDU (Martin Schoffstall)
Wed Aug 28 21:53:48 1991
In-Reply-To: <CMM.0.90.2.683330294.vaf@Valinor.Stanford.EDU>
Date: Wed, 28 Aug 91 20:30:50 -0400
To: "Vince Fuller" <vaf@valinor.stanford.edu>
Cc: "William Schrader" <wls@psi.com>, ittai@shemesh.ans.net,
From: "Martin Schoffstall" <schoff@mail.psi.net>
Reply-To: schoff@psi.com
>DATE: Tue, 27 Aug 91 14:58:14 PDT
>FROM: Vince Fuller <vaf@valinor.stanford.edu>
>
>Excerpting from:
>
>This is an interesting statement. From a legal and political point of view,
>this may be true. From the technical standpoint, however, the existance of
>both CIX and ANS serves to exacerbate the scaling problems associated with IP
>routing - with multiple external attachments, a mid-level may need to carry
>the global routing database; with one attachement, it need only carry its
>own routes plus a default route to the external location. Maybe IDPR and/or
>Noel's new IP scheme will save us all in the end, but I'm expecting the worst
>in the meantime...
It does not exacerbate the scaling problems for standard end-user
connections that 99.99% of internet customers have - because they have a
single connection, and almost all mid-levels have hid all complexities
of routing from them from day one.
And I concur with you that it doesn't exacerbate the scaling problem for
a mid-level with a single attachment. But how many of those were there
out there before the CIX? Let's do a survey:
Network Connections
------- -----------
SURANET NSFNet,NSFNet,ESNET,MILNET
PSINet NSFNet,TWBNet,ILAN,XLINK,DND
JVNCNet NSFNet,RISQ,BITNET2-TCP,.....
CERFNet NSFNet,LosNettos,ESNet,Korea....
NEARNet NSFNet,ESNet,ALTERNET
ALTERNET NEARNet,NSFNet,EUNET..
Now I apologize for letting lots of things out and making mistakes, but
you get the idea.
Now let's take a look at the case of a regional built like a star with a
NSFNet connection. It is easy for the hub of the tree to gain knowledge
from the NSFNet about its network numbers, talk to the CIX for its
network numbers, and hide all of that from the subnet and from the customers.
Peace of cake.
I'm looking to Noel to solve part of this problem too, but we have the
pieces today to do this. In fact what do you really think is going on
inside those UNIX timesharing machines inside the NSFNet?
Marty
