as1275

Pedro Marques roque@cisco.com
Mon, 20 Oct 1997 11:14:12 -0700 (PDT)


>>>>> "Guido" == JOIN Project Team <ipng@uni-muenster.de> writes:

    Guido> Hi Pedro,

    Guido> On Sun, 19 Oct 1997, Pedro Marques wrote:

    Guido> AS1275 is registered (in the RIPE db) as "DFN-IP service
    Guido> and DFN customer networks". We use it for our IPv6 JOIN
    Guido> project (it is an DFN project).

There should be only one entry for as1275 in the 6bone registry. That would
make figuring out the right site much easier.

    >> What i don't really understand is why, if all the sites share
    >> the same AS, aren't they all under the same prefix which can be
    >> cleanly annouced. And if there is a good reason why this can't
    >> be done then perhaps they shoulnd't be in the same AS...

    Guido> All IPv6 sites using AS1275 should have the same prefixes
    Guido> 3ffe:400::/24 or 5f04:fb00::/32. In general these are
    Guido> German sites like Universities.

Good.

    Guido> For these two prefixes we do BGP route aggregation, so i
    Guido> wonder if you see more and/or longer prefixes *originated*
    Guido> at JOIN.

Like i mentioned in my previous message i see close to the full routing
table *originating* at as1275.

    Guido> These days i discussed similar problems with Guy
    Guido> (UUNET-UK). He also received a lot of routes which should
    Guido> be originated by JOIN according to his routing table. Now
    Guido> this seems to be the case also at your site :-(

    Guido> When i examine our routing table i find for all the routes
    Guido> (which you claim as "tons of doubtable information"), a
    Guido> source from which our router gets them.

The important thing would be to figure out what your router annouces. Either
misconfiguration or software bug, it would be nice to figure out the problem.

    Guido> I wonder why your routing table mentions AS1275 as
    Guido> originator of all these routes.

We have no problems whatsoever with routes from any other sites. I pretty
much doubt those path are "invented" along the way.

    Guido> What does "i" (IGP) as origin code mean in your routing
    Guido> table?
It is the value of the BGP origin attribute (check rfc 1771).

    Guido>   Does it mean that these routes are learned via
    Guido> RIPng at our site?

The value of that attribute is set by the originator of a bgp annoucement.

    Guido>   I believe there should be "e"'s (EGP),
    Guido> because we got these routes via BGP from other sites.

I believe that is not too relevant at the moment. The important thing is to
figure out how to solve the problem.
 
    Guido> Some of your routes specify "?" (incomplete) as source of
    Guido> (our?)  information: when i check these routes i get
    Guido> "BGP-INCOMPLETE" in our routing table. That means, that
    Guido> this route is learned via BGP from our router, but has not
    Guido> a complete BGP path to the originator of the prefix -
    Guido> e.g. there is some RIPng or static routing inbetween.

No, it doesn't mean anything close to that.

    Guido> I did not find a default route in our table. Maybe this
    Guido> route was visible at the time you examined your table - but
    Guido> i am quite sure that we are *not* the originator of a
    Guido> default route (hopefully i am right ;-)

I really hate to disappoint you but that seems to me to be the most probable
alternative at the moment. The others would be for the AS path you are
annoucing to get truncated by one of your peers, or having someone else
using your as #.

  Pedro.