[rrg] The case for NAT66

[Christian Vogt]

Iljitsch -

You are right.  Fail-over functionality for multi-homed networks suffers
from a deployment problem because it requires support on both ends of a
connection.  We could conquer the deployment problem if we specified the
fail-over functionality as part of NAT66, thereby piggybacking the
deployment of the fail-over functionality to the deployment of NAT66.

In fact, in doing so, we would exploit vendors' desire for standard
compliance to make them implement the fail-over functionality, and we
would exploit operators' perceived benefits of NAT to make them deploy
the functionality.  Sounds like a reasonable deployment strategy.

BTW, we did already discuss this idea on the BEHAVE list.  Check this:

http://www.nabble.com/Why-IETF-should-standardize-IPv6-NAT-to20290669.html#a20293096

In that thread, we considered that NAT66 is the same as the Unilateral
mode of Six/One Router [1], and that we could augment NAT66 to support
also Six/One Router's Bilateral mode.  Having said this, I do believe
that NAT66 should be standardized also for other reasons [2].

- Christian


[1] http://users.piuha.net/chvogt/pub/2008/vogt-2008-six-one-router-design.pdf

[2] http://www.ietf.org/mail-archive/web/behave/current/msg04611.html