North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: Bad IPv6 connectivity or why not to announce more specifics(Was: IPv6 news)

  • From: Jeroen Massar
  • Date: Thu Oct 13 15:49:39 2005

On Thu, 2005-10-13 at 15:38 -0400, [email protected] wrote:
> On Thu, 13 Oct 2005 21:06:30 +0200, Jeroen Massar said:
> > Well Valdis, that bad route also has to do with your side of the
> > equation, you might want to check who you are actually using as transits
> > and if the routes they are providing to you are sane enough.
> Well, if somebody at wants to do a traceroute6 back at us, I'll
> be glad to see what the reverse path looks like...  but last I heard
> traceroute and traceroute6 showed the *forward* path of packets..

That is correct, try tracepath, this shows at least the assymetry.
You can also peek at GRH to see a probable AS path back. ASN's still
tell a lot in IPv6.

Next month I'll finalize the 'symmetry' tool which allows one to do the
AS path checkup between two places automatically.

> > 2001:468::/32 is in the routing table, getting accepted by most ISP's.
> > This one has a reasonable route
> The real problem (at least for the forward direction from here) is that the
> outbound packets get into the Abilene network, and the best path from there to
> 2001:440:1880 is a 3ffe: tunnel to japan and then another 3ffe: tunnel back to New
> York.

Kick Abilene to not be so silly and get some real transits. Then again
Abiline is educational and those networks seem to have very nice (read:
overcomplex) routing policies...


Attachment: signature.asc
Description: This is a digitally signed message part