[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Unaggregated prefixes in the BGP4+ cloud

Hi all,

as many of you probably already know, at CSELT we are making available a
set of html pages (http://carmen.cselt.it/ipv6/bgp/index.html) showing
the BGP4+ routing tree from CSELT to the whole 6Bone and from CSELT to
the 6Bone backbone.
These pages are updated every 5 min and looking at the most recent
results (http://carmen.cselt.it/ipv6/bgp/bgp-page-complete.html) I have
observed that I am receiving several BGP4+ routes towards excessive long
prefixes (/124 and longer!). In addition some of these routes are
associated to excessive long AS paths (from 10 to 15 hops!).

I could easily solve the problem placing some BGP4+ input filters on my
IPv6 border routers (e.g. blocking all the prefixes longer than /64
should be enough to clean up my routing tables).
But before doing that I would like to know if somebody else is already
adopting this kind of policy inside the 6Bone.

I am wondering if the correct way to solve the problem is to place
filters on the routers or to send e-mails in order to help every BGP4+
speaker to perform correct advertisements.

Opinions would be appreciated.



Ivano Guardini  CSELT SpA
via G. Reiss Romoli 274  Torino (Italy)
Tel. +39 11 228 5424
Fax. +39 11 228 5069
e-mail: [email protected]