draft-durand-v6ops-dualstack-vs-natpt-00

["Bound, Jim" <Jim.Bound@hp.com>]

I think this draft is important for the working group to work on and
deal with the issues in this work.

--------------------------------------------
Abstract

   Outside of the IETF community, lot of people think that IPv4 to IPv6
   transition consist merely at solving the problem of how does a v4 box
   communicate with a v6 box and vice versa.  Within the IETF, the dual
   stack approach has long been defined.  There is an ongoing discussion
   to understand if translation with tools like [NAT-PT] is absolutly
   needed to enable IPv6 nodes to communicate with an IPv4 node or if we
   can/should mandate IPv6 nodes to also deploy an IPv4 stack if/when
   they needs to communicate with IPv4 nodes.  This draft is aimed at
   clarifying the discussion without taking side by studying in 3 cases
   the implications of mandating a dual-stack versus the implications of
   deploying a translation device
----------------------------------

I don't like the idea of mandating anything that's a bad word above in
the last sentence.  But this is an area that should be documented so
operators understand the ramification of NAT with IPv6 and the
trade-offs.

Most of my input are edits I will send offline and still thinking if
this is enough cases, but I think so.

It is stated all understand the dual stack concept but I do not believe
entirely.  For example IPv6 don't run over IPv4 as one error I have
heard regarding implementation.  Also most implementations are really
dual IP layers not dual stacks.  They are more of what is called a
"hyrid IPv4/IPv6 stack".  

But this work should be pursued.

/jim