North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: peering wars revisited? PSI vs Exodus

  • From: Valdis.Kletnieks
  • Date: Tue Apr 04 00:39:36 2000

On Mon, 03 Apr 2000 21:47:50 EDT, Paul Ferguson said:
> At 09:27 PM 04/03/2000 -0400, Gordon Cook wrote:
> >surprised not to see this mentioned on NANOG
> > >Sent:  Friday, March 31, 2000
> > >To:    Notify
> > >Subject:       Exodus Customer Confidential Communication
> Gordon,
> Does the word "confidential" elude you?
> - paul

In Gordon's defense (I feel like the ACLU here ;),

(a) Yes, professional courtesy does usually respect a "confidential"
label.  However, you have to remember that Gordon has been acting as a
whistleblower for quite some time, and if Gordon hadn't posted it,
there'd be a lot of people pondering the lack of BGP announcements and
labelling it "just another screw-up" rather than a actual deliberate

(b) By the time Gordon posted it, the announced disconnect time had
already passed by several days.

(c) I am *hoping* that the Exodus message was mistaken when it said
that "*all* connectivity" would be lost - PSI isn't *really* dropping
all packets on the floor if they happen to have come from an
Exodus-connected network, even if they are arriving via a PSI-Sprint or
whatever peering?

If it's merely "PSI refuses to peer directly with Exodus", that's one
thing.  If PSI is also refusing packets carried by some other 3rd party
that PSI and Exodus both peer with, or alternate routing is failing
for some other reason, that's a lot worse.

Does anybody know definitively what the REAL story is?

				Valdis Kletnieks
				Operating Systems Analyst
				Virginia Tech