Hi Suresh,
Could be all of what you suggest... I've really not had the time to go
delving into these (plus it would need the willing cooperation of the AS
in question, which might have happened 25 years ago when some of us were
chasing down meaningless deaggregation). These massive only affect
certain paths and, in RouteViews case, certain peers in certain
locations (some locations they send us vast numbers of updates, others
they are perfectly normal). I could speculate infrastructure issues, or
internal routers failing to converge, or BGP attribute changes, or a
script gone mad, or broken BGP implementations,... Not sure if it is
those automatic load balancers tbh, they don't show the vast level of
updates, but just generate a steady background noise (you can see the
obvious candidates in Geoff's BGP Update Report).
philip
--
Suresh Ramasubramanian wrote on 10/2/2025 12:16:
Philip, I’ve kept meaning to ask you for decades but did you ever find
the root cause of all this noise- is this due to a bug in some
version of a router OS or is this being caused by a malfunctioning
script at the provider that triggers BGP updates?
From: NANOG <nanog-bounces+ops.lists=gmail....@nanog.org> on behalf of
Philip Smith <p...@routeviews.org>
Date: Monday, 10 February 2025 at 6:32 AM
To: James Bensley <lists+na...@bensley.me>
Cc: NANOG <nanog@nanog.org>
Subject: Re: Noisy prefixes in BGP
I guess all we can do is keep highlighting the problem (I highlight
Geoff's BGP Update report almost every BGP Best Practice training I run
here in AsiaPac, for example) - but how to make noisy peers go away,
longer term...? :-(
> * These problems aren’t DFZ wide. Peer A might be sending a bajillion updates to peer B, but
peer B sees there is no change in the route and correctly doesn’t
forward the update onwards to it’s peers / public collectors. So this
is probably happing a lot more than we see via RIS or RouteViews. Only
some parts of the DFZ will be receiving the gratuitous updates/withdraws.
