[193324] in North American Network Operators' Group
Re: Soliciting your opinions on Internet routing: A survey on BGP
daemon@ATHENA.MIT.EDU (Laurent Vanbever)
Tue Jan 10 19:06:04 2017
X-Original-To: nanog@nanog.org
From: Laurent Vanbever <lvanbever@ethz.ch>
Date: Tue, 10 Jan 2017 08:13:51 +0100
In-Reply-To: <6260f07f-083d-f88c-4656-1a0d14e7aee8@bogus.com>
To: joel jaeggli <joelja@bogus.com>
Cc: NANOG list <nanog@nanog.org>
Errors-To: nanog-bounces@nanog.org
Hi Joel,
> On 10 Jan 2017, at 06:51, joel jaeggli <joelja@bogus.com> wrote:
>=20
> On 1/9/17 2:56 PM, Laurent Vanbever wrote:
>> Hi NANOG,
>>=20
>> We often read that the Internet (i.e. BGP) is "slow to converge". But =
how slow
>> is it really? Do you care anyway? And can we (researchers) do =
anything about it?
>> Please help us out to find out by answering our short anonymous =
survey=20
>> (<10 minutes).
>>=20
>> Survey URL: https://goo.gl/forms/JZd2CK0EFpCk0c272 =
<https://goo.gl/forms/WW7KX5kT45m6UUM82>
>>=20
>>=20
>> ** Background:
>>=20
>> While existing fast-reroute mechanisms enable sub-second convergence =
upon=20
>> local outages (planned or not), they do not apply to remote outages =
happening=20
>> further away from your AS as their detection and protection =
mechanisms only=20
>> work locally.
>>=20
>> Remote outages therefore mandate a "BGP-only" convergence which tends =
to be
>> slow, as long streams of BGP UPDATEs (containing up to 100,000s of =
them) must
>> be propagated router-by-router. Our initial measurements indicate =
that it can
>> take state-of-the-art BGP routers dozens of seconds to process and =
propagate
>> these large streams of BGP UPDATEs. During this time, traffic for =
important
>> destinations can be lost.
>=20
> One of the phenomena that is relatively easy to observe by withdrawing =
a
> prefix entirely is the convergence towards longer and longer AS paths
> until the route disappears entirely. that is providers that are =
further
> away will remain advertising the route and in the interim their
> neighbors will ingest the available path will until they too process
> the withdraw. it can take a comically long time (like 5 minutes) to =
see
> the prefix ultimately disappear from the internet. When withdrawing a
> prefix from a peer with which you have a single adjacency this can
> easily happens in miniature.
Thanks! Yes, definitely. This relates to the issue Baldur was raising in =
which a less-preferred prefix (or not prefix at all in your case) has to =
take over a more preferred one. That case is definitely bad for BGP =
convergence.=20
Our survey/study is more geared towards cases where there is diversity =
available (alternates paths are there and at least partially visible). =
We are especially interested in finding out whether, even when you take =
all the precautionary measures required by the book, long BGP =
convergence can still bite you and=E2=80=A6 whether we can do anything =
about it.
Laurent
PS:=20
Thanks so much to the 21 operators who have answered already! If you =
haven=E2=80=99t so already, please help us out to find out about =
troublesome BGP convergence by answering our short anonymous survey =
(<10 minutes): https://goo.gl/forms/JZd2CK0EFpCk0c272 =
<https://goo.gl/forms/JZd2CK0EFpCk0c272>
