[481] in Commercialization & Privatization of the Internet
Re: CIX Implications
daemon@ATHENA.MIT.EDU (Kent England)
Fri Mar 29 09:28:06 1991
From: kwe@bu-it.bu.edu (Kent England)
Date: Fri, 29 Mar 91 09:26:50 -0500
To: com-priv@psi.com
In-Reply-To: Mail from 'oleary@sura.net (dave o'leary)' dated Thu, 28 Mar 91 19:27:57 -0500
Cc: oleary@sura.net
Dave O'Leary notes:
>
> Currently AlterNet and SURAnet exchange routing information at the SURAnet
> NOC in College Park. I don't know if AlterNet is sending us all their
> routes, or just for those networks that are "NSFnet approved". I don't know
> if they intend to send us CERFnet and PSInet routes.
>
This points up the difficulty of global co-ordination of
routing as *IXs proliferate.
> So presumably, if Harris (or any of SURAnet's other customers) want to
> send or receive commercial traffic from any sites on AlterNet, they
> can already do so. I don't know if AlterNet is willing to serve as a
> transit network for SURAnet <-> PSInet traffic (for example).
>
What Dave doesn't say but would certainly point out when it
came up, is that the CIX/FIX and all other Internet routing works on a
"per net" basis with independent transit net selection at each end; so
if a SURAnet site chooses to send commercial traffic to someone on
Alternet via the CIX, it may very well come back via the NSFnet,
unless they explicitly tell the NSFnet world to please classify their
net as "no NSFnet transit" and to route via Alternet/CIX or some other
unrestricted backbone that they have access to from their side.
The CIX is not the answer to the problem of how one net sends
and receives both restricted and unrestricted traffic over multiple
backbones. That problem still remains unaddressed by currently
deployed technology.
--Kent
