On killing IPv6 transition mechanisms

Tore Anderson tore.anderson at redpill-linpro.com
Thu Mar 18 23:58:16 CET 2010


* Brian E Carpenter

> I believe this was intentional, since the philosophy was 'prefer IPv6
> connectivity if it exists'. I also agree that reversing this in
> the policy table would be very reasonable; 3484 only provides a
> *default* policy table.

I should probably have said "shortcoming" instead of "bug".  However,
the RFC does special-case 6to4 and makes 6to4-to-!6to4 connectivity less
preferred than IPv4-to-IPv4.  If the point was to prefer IPv6
unconditionally, they would not have had to mention the 6to4 prefix at all.

But they did, and I cannot imagine any other reason that they recognized
that 6to4-based connectivity would be inherently less reliable than the
IPv4 connectivity that 6to4 was tunneled on top of. So I don't quite
understand why they didn't make 6to4-to-6to4 less preferred than
IPv4-to-IPv4 too.  I don't think 6to4-to-6to4 is used for much
production traffic though, so it's not really a problem.

> You mean, I assume, that it turns out that there is no 6to4 relay
> in the outbound or return path? That is (IMHO) the problem
> with RFC 3068. The assumption of RFC 3056 was that the placement
> of 6to4 relays would be a managed process, for savvy early-adopter
> sites. The addition of the anycast relay model in RFC 3068 broke that
> assumption and brought about unmanaged deployment.

Yes, that there is no 6to4 relay in the outbound path (I can easily make
sure there's return relays for 2002::/16 within my own network so that
part isn't easily handled), or that proto-41 traffic is filtered, or
that the relay doesn't work, or is very far away latency-wise, and so on
and so on.

Broken 6to4/Teredo connectivity is the single largest cause for web
sites to appear unreachable/down for regular eyeballs after introducing
AAAA records.  It's responsible for almost all of it, I cannot really
point out any other single cause of significance, of the 0.094% of total
brokenness I measured in February, 0.090 percentage points is due to

Had it not been for 6to4 I doubt I would have had an problems convincing
my customers to dualstack their content.  But now they're understandably
concerned about losing visitors and thus revenue. Probably most content
providers wants everyone else to go first and take the pain, so that the
problematic users hopefully will be mostly fixed by the time they add
IPv6 capability themselves.

> Nathan Ward has suggested a way to fix this (and the equivalent
> problem with Vista/Teredo) but it does need sites or ISPs to install
> a box to catch the anycasts.

Do you have a pointer to his suggested solution?

Best regards,
Tore Anderson
Redpill Linpro AS - http://www.redpill-linpro.com/
Tel: +47 21 54 41 27

More information about the ipv6-ops mailing list