Re: successful termination?

[Christian Vogt <christian.vogt@ericsson.com>]

Dear all -

I apologize for getting involved in this interesting discussion only
now.  A few late comments from my side regarding the potential
continuation of the Shim6 working group:

In my opinion, this discussion has shown that there would be useful work
to do for a continued Shim6 working group.  In particular Marcelo's
proposal to evolve Shim6 as a locator-pair-providing substrate for
multi-path TCP is a very good idea.  One may even go a step further and
provide a similar service also for UDP-based applications.

A second potential future work topic is adding capabilities for network-
controlled traffic engineering to Shim6.  The (host-based) Six/One
protocol [1] could serve as a starting point for this.  Erik Nordmark,
too, had had ideas [2] for network-controlled traffic engineering.

[1] http://tools.ietf.org/html/draft-vogt-rrg-six-one-01
[2] http://tools.ietf.org/html/draft-nordmark-shim6-esd-01

A third potential future work topic is deployment facilitators for
Shim6.  This would include, in particular, a network-based variant of
Shim6, which would have at least the following two benefits:  First, it
could aid Shim6 adoption during the phase when (host-based) Shim6 is not
yet widely deployed.  Second, it would provide an option to keep multi-
homing control in the network, a feature that some network operators
apparently prefer as the discussion in early RRG days had shown.

One approach to design a network-based Shim6 variant would be to build
on Proxy Shim6 [3].  An alternative approach could be to build on
(network-based) Six/One Router [4].  As for the latter, it might be
possible to borrow techniques developed elsewhere in IETF:  As you may
know, the recently discussed NAT66 proposal is identical to Six/One
Router when used on only one side of a connection.  What is missing is
support for reverse translation when Six/One Router is used on both
sides of a connection, plus a distribution system for address mappings.
Support for reverse translation could be specified as an extension to
NAT66. Support for mapping distribution could be done as in Proxy Shim6.

[3] http://tools.ietf.org/html/draft-bagnulo-pshim6-02
[4] http://christianvogt.mailup.net/pub/2008/vogt-2008-six-one-router-design.pdf

- Christian