Survey on Internet Disputes.

Mailman List Mirror (Read Only)
1

Hi,

I am Kshitiz Verma, a Ph.D. student in Madrid, Spain, working in the area
of studies on the Internet ecosystem. We are conducting a survey on the
Internet disputes, not limited to, but mainly focusing on de-peering. We
will appreciate responses from the community that help us build our data on
such incidences. We did an initial search (
https://docs.google.com/spreadsheet/ccc?key=0Ajn7QWBNzSlLdGo3UzFDVkNvdGVwYncxNUdJclZWcnc&usp=sharing#gid=0),
but unfortunately the results we obtained were mostly limited to well known
disputes in North America. We believe there should be more of them but are
hard to find,
1. if they occur outside the Tier 1 regime.
2. if outside USA.
3. because the local news may not always be in English.
4. they may not last long or may not have as big an impact and find place
in news.

We are posting We are also looking for instances like Youtube hijacking by
Pakistan Telecom is not a dispute between any two ASes but still for
repairing purposes, connectivity between Pakistan Telecom and PCCW (AS
3491), Hong Kong was disrupted. At the same time, we are not looking for
disruption of the Internet connectivity due to an accident like Taiwan
earthquake in December 2006.

As claimed in http://www.cs.columbia.edu/~misra/news/CD070113.pdf , 500 to
1000 de-peering happens on a daily basis today. We are curious to know how
this number has evolved since its inception (starting from post
commercialization of NSFNET is already very good but before that is
excellent :slight_smile: ).

Please fill in the following form if you have some information that you may
want to share with us. Also, as there are so many de-peerings happening,
there is a high possibility that most of these don't become International
News, so to get information on these disputes we expanded our search to
local and regional News but unfortunately this has not added significant
information. Once the results are processed and filtered, we intend to
share them publicly.

Form:
https://docs.google.com/forms/d/1X6msZeK0_3SIaIK8IxZaO2-Vrl1MPh_8dTSpbG0oGzY/viewform

Many thanks in advance :slight_smile:
Kshitiz Verma

2

I suppose this is just by technical incapabilities. People leak
prefixes, hit max-pref limters, forget to clear sessions or don't bother
increasing limits, they migrate gear from Cisco to Brocade or Juniper
and cannot recover encrypted MD5 passwords... or management decisions
decide to pull from an exchange.

I don't consider this "de-peering" a peering dispute or an offensive
act. The majority of vanished BGP adjacencies are due to laziness or
technical limitations.

3

Thanks for the clarification on the number. I was surprised to see that
number too!

At the same time, we couldn't even find genuine disputes apart from the
ones we shared. It seems there should be more but we just could not find
them on the web.

4

Much more common than actual depeering is the passive-agressive version,
where you continue to peer but bring a smaller hose to the peering point
than the peering partner does. We see a *lot* of reports of the resulting
congestion and packet misbehavior on this list...