AT&T via Tata and Level3

Robert Bonomi bonomi at
Thu Mar 3 17:39:06 CST 2011

> Date: Thu, 3 Mar 2011 14:12:16 -0600
> From: Richard A Steenbergen <ras at>
> Subject: Re: AT&T via Tata and Level3
> Cc: nanog at
> On Thu, Mar 03, 2011 at 11:15:51AM -0500, Morgan Miskell wrote:
> > I've noticed that we have thousands of routes for AT&T via Tata that we 
> > don't have from AT&T through Level3.  I would expect Level3 to have 
> > most of the routes for AT&T that Tata does since they are both directly 
> > peered with AT&T.
> Well, I don't know anything about this specific issue or any policy 
> changes that may have been made, but at a high level I can tell you that 
> BGP doesn't work like that. BGP is only capable of passing on a single 
> best path for each route, and what is considered the best path is totally 
> in the eye of the beholder.

[[..  sneck much good stuff  ..]]

While what you say is accurate, it is _irrelevant_ to the situation that
the OP posted about.  Methinks you misunderstood what he said.

He peers with Level3 and TATA.  Both of whom peer with AT&T.

Looking at the -incoming- data from those two peers, he sees "thousands"
of entries for AT&T address-blocks announced to him by TATA that are 
not being announced to him by Level3.

Postulating that AT&T _is_ announcing all its address-blocks to both of 
those direct peers, the 'one-BGP-hop-removed-from-directly-connected' 
network should expect to see all those blocks from any of it's directly 
connected peers that are directly connected to AT&T.  If one of those 
peers sees a 'better' route to one of those AT&T address-blocks, then it
should be announcing that indirect path instead of the direct one.  Ditto
for blocks that AT&D does -not- announce (for whatever reason, traffic
engineering, maybe?) to a directly connected peer.

I would hazard a guess that the "missing routes" _might_ be the result of
supressing 'more specifics', or they _are_ being announced to Level3, but
with a 'community' tag that Level3 interprets as 'use locally, but do not
announce externally'.

More information about the NANOG mailing list