Say "Thank you" to Bill...

Max Tulyev maxtul at netassist.kiev.ua
Sun Apr 1 16:08:30 CEST 2007


Bernhard Schmidt wrote:
>> teredo.ipv6.microsoft.com. 3600 IN      A       65.54.227.136
> Please have a look how Teredo _servers_ (which is what Microsoft^WBill
> is offering) are involved in the communication.
> 
> http://www.microsoft.com/technet/prodtechnol/winxppro/maintain/teredo.mspx
> http://www.ietf.org/rfc/rfc4380.txt
> 
> Hint: Teredo Server != Teredo Relay, the latter are operated by a number
> of people, but not Microsoft.

Thank you, nice reading ;)
Am I understand correctly Server helps to chose a Relay?

> We have enabled IPv6 on our webfarm, so far no complaints. I think we
> are one of the few users of the IPv6 Gateway thing on F5 BigIP
> loadbalancers. But of course, we have decent upstream for that. I reach
> you through
> 
> M-net -> Verio -> ISC -> NetAssist

I know, we are working to fix it (i.e. do some traffic engineering, set
prepends etc to see the world with shortest way.

We are going to establish layer2 channel to Frankfurt to get native and
shortest V6 instead of tunnels. It takes some time, of course.

> I would not dare to run any "interactive services" (I consider
> webbrowsing interactive) on that either.

Browsing working fine, by the way. SSH have a significant lags.

> I still cannot believe that this is actually a problem of the end-site
> client. Opposed to 6to4, where you indeed crash into firewalls all the
> time, Teredo has a qualification mechanism. If you don't reach the
> Teredo server, you won't get an address, and Teredo won't be used. It
> can still be broken, but it is hard to break real Teredo traffic
> unintentionally when Teredo qualification worked.

We are continue investigation. Only is clear - disabling IPv6 in hosting
stops complains and gets sites back visible, and this happens with Vista
and XP enabled V6 clients.
The second thing may be MTU issues, so there is link, all reachable, but
for example fotos are not loading.

-- 
WBR,
Max Tulyev (MT6561-RIPE, 2:463/253 at FIDO)


More information about the ipv6-ops mailing list