[6bone] 3ffe:3400::/24 / IPF/DE

[Bob Fink]

Jeroen,

Thanks for your help on this.


Bob

===
At 12:55 PM 12/4/2002 +0100, Jeroen Massar wrote:
>Bob Fink wrote:
>
> > At 08:14 PM 12/3/2002 +0100, Pim van Pelt wrote:
>
> > >I think there's quite some (nearly) dead pTLA deployments out there.
> > >
> > >A periodic check might not be that bad either. There's plenty of
> > >public LGs around to confirm the absence of a prefix. Also,  RIPE
> > >has just started the RISv6 project (I was their first peer, awww :)
> > >and we might be able to make use of that also, to detect things.
> >
> > It would be great to have some automatically reporting tool
> > that we could
> > use to evaluate what is no longer in use.
>
>I was already collecting prefixes from the bgp dumps I got for the
>Ghost Route Hunter* project. I'll grab the current pTLA list and
>make a report page for comparing availability per pTLA along with
>their origin AS. This also allows us to easily verify if the 6bone
>object for that prefix corresponds with the AS it originaties from.
>Ghost routes and misannouncements in there will also pop up :)
>
>Notez bien that Gert Doering probably has a much bigger archive
>due to his IPv6 routing overview's (www.space.net/~gert/RIPE/)
>And that the ASPath tool can also quite clearly show which TLA's
>can be found where.
>
>The RISv6 tool unfortunatly only carries BGP entries from native peers.
>So most parts of the 6bone space are probably not visible there. RIS is
>more extensive than any other routing project though
>and they keep histories and other relevant information like updates.
>So it could proof very usefull if many ISP's hook up with it.
>
>8<----------------
>Found 19743 no, 0 short, 21 long ghosts.
>Longest: 31
>Prefixes: 436
>--------------->8
>The prefixes there also hold /48's and others which are localy
>visible, but I'll filter those out for the pTLA list ofcourse.
>The first line shows ASpaths btw. Doing a rough calculation
>this would mean I got 45 paths for every prefix. Skimming through
>it quickly reveals that most TLA's have about 65 paths available.
>Spread over 9 BGP feeds, more are always welcome ofcourse.
>
> > I am starting to request pTLA holders that I become aware are
> > obviously not
> > using their pTLA for longer periods of time (months) to give
> > back their
> > pTLA and reapply for a new one if they want another in the
> > future (i.e.,
> > not just hold on until they change their plans).
> >
> > Sounds like 3FFE:3400::/24 is a strong candidate to be
> > reclaimed. I will
> > post an intent to reclaim this pTLA with a two week notice.
>
>Reclaiming is good. This will limit problems and we will be able
>to identify them much better. But the availability of an AS
>doesn't have to mean that the prefixes are available.
>DIGITAL-CA for example _was_ only available to the 'old' prefixes.
>eg 3ffe:3000::/24 (AMS-IX) could reach it, or better said they
>knew where to find AMS-IX, but they didn't knew the way back to RIR
>space. This can be seen in the https://www.sixxs.net/misc/latency/
>graphs looking at the host.ipv6.pa-x.dec.com router.
>But apparently it started pinging for Cybernet, Concepts and Intouch
>now... so the upgrade worked out as they told me when I queried them.
>
>Also IMHO 6bone should move to the RIR's ASAP.
>(RIR space are paying people, no free transit -> better stability
>blablabla, see those other threads :)
>
>Awareness is a good thing.
>
>Greets,
>  Jeroen
>
>* = https://www.sixxs.net/tools/grh/
>
>_______________________________________________
>6bone mailing list
>6bone@mailman.isi.edu
>http://mailman.isi.edu/mailman/listinfo/6bone