RE: 6to4 public anycast relay considered a bad think (was Re: 6to4 connectivity test)

[Christian Huitema <huitema@windows.microsoft.com>]

> Essentially, expecting to get a functioning 6to4 relay is expecting a
> free lunch. Who is going to pay for it?

Alain is using some extreme language, but there really is an issue with the deployment model for 6to4 relays.

James is observing that some public relays are broken, perhaps deliberately. This point was raised during the discussion of the "anycast relay" RFC. It is actually a well known failure mode of the anycast model: your traffic is just a routing update away from a black hole. The remedy is well known: 6to4 routers ought to be configurable. If the user can somehow procure a reliable gateway, that gateway should be used.

Teredo went one step further. Public gateways can easily be abused. So Teredo introduced a discovery mechanism to find out the best gateway on a destination by destination basis. That mechanism is little more than a ping, and could easily be ported to 6to4. We could assume that 6to4 routers maintain a "routing cache" associating specific "native IPv6" destinations with the "closest 6to4 gateway". Given a new IPv6 destination, the 6to4 router will send a ping through the public server, note the IPv4 address from which the ping comes back, and send the rest of the traffic through that address.

In short, the Teredo model places the deployment onus on the IPv6 only servers. They must somehow ensure presence of a transition relay near them. They have an incentive to do so: serving the users of the transition technology. They also have an incentive to provide good quality service. This is a big contrast with the 6to4 model, which basically places the requirement on the legacy ISP.

-- Christian Huitema