Re: [RRG] Conceptual vs. specific - another discussion list?

[Robin Whittle <rw@firstpr.com.au>]

Hi Tony,

Thanks very much for offering to set up another IRTF list.

I would like to keep this thread going to discuss the suggested
scope, charter or whatever.  If you don't want that, then I would
appreciate people writing to me offlist so I can coordinate their
desires into a proposal and send it to you directly.

You wrote, in part:

> Not at all.  As we stated, solutions that apply for v4 and v6 are
> certainly within scope.

OK, but I have two concerns:

  1 - You have several times tried to steer discussion away from
      specific proposals.

  2 - If the proponents of such solutions are primarily interested
      in applying their solution to IPv4 in a shorter time-scale
      than the RRG consensus view indicates is necessary (the new
      rough consensus indicates the RRG may decide no IPv4 solution
      is ever necessary) then this discussion will get in the way
      of the higher-level conceptual discussion I think you want
      the RRG to focus on.


>> Likewise, does this mean that discussions about migration of
>> large numbers of users to IPv6, or to a future significantly
>> modified version of IPv6, are out of scope?  My guess is they
>> would be, and would disrupt the discussions you want to focus
>> on.
> 
> Yes, you're correct.  We've spent 15 years talking about v6
> migration and there's been very little light shed for the last 14
> of those.  ;-)

OK.  Discussion of IPv6 migration strategies, IPv4 address
reclamation, time-frames for IPv4 meltdown and IPv6 blossoming etc.
is clearly off-topic for the RRG list, but would be appropriate on
the new list.

I think the folks who don't share every consensus view still have a
lot to contribute to the overall process of developing a new
Internet architecture - and would like to discuss this in close
association with the RRG.


> I've got no problems moderating more lists and if folks want to
> spin off subgroups to go have disucssions that's also welcome.
> I'd like to know what subjects/problems/projects/experiments they
> are going to deal with.

That is very generous.


> I interpret research as simply broader than engineering.
> Engineering is the art of building based on the knowledge at
> hand.  Research also encompasses pushing back the boundaries of
> knowledge.

I think "research" is developing new knowledge about the natural
world and/or human constructed systems. I have no problem with
thinking of our work as being research into engineering principles
and techniques.  This may involve new or existing scientific or
mathematical principles.

I think the goal is to artfully combine known and perhaps novel
mathematical and engineering techniques and principles in order to
create the optimal architecture.  However, we also need to think
about existing practice, commercial considerations, the needs of
end-users etc.  Otherwise, no-one will adopt what we devise.


> Since what we're doing is computer networking, which is largely
> based on graph theory, complexity theory, computer scient, and
> topology, I'd say that what we're doing is definitely applied
> mathematics.  Therefore it counts as hard science.  I have to
> admit bias here, as I'm trained as a mathematician and computer
> scientist.

Maybe some computer networking architectures suffered from too
little focus on mathematically purity and therefore couldn't scale
so well as to operate on a global scale.

Perhaps overdoing the mathematical focus to the exclusion of other
considerations could lead to an architecture which is insufficiently
backwards compatible with existing practice to be widely adopted.


Here is a rough proposal, written as if there are to be more than
one moderator (which I don't assume) and assuming that you are one
of those moderators.

  The new list might be called:

     RRG-Sandbox

  [RSB], [RRGS] or [RRGSB] might be the subject header - the shorter
  the better.

  Membership of RRG-Sandbox is expected to be a subset of the RRG
  membership.

  All actively posting RRG-Sandbox members are expected to do so
  in awareness of current and recent RRG discussions.

  RRG-only members are not expected to take an interest in
  RRG-Sandbox.

  RRG-Sandbox is not intended as a venue for forming consensus of
  any kind.  The question of RRG consensus is determined solely by
  the RRG Moderators based on RRG messages.

  The Moderators may direct an RRG-Sandbox thread to be continued on
  the RRG list - or an RRG thread to be continued on RRG-Sandbox.

  Relevant topics include:

     New and modified Internet architectures concerning
     scalable approaches to routing and addressing, for
     purposes including:

        Multihoming for end-user networks.

        Traffic engineering.

        Portability or other approaches to easing the
        switch to another ISP.

        Mobility of networks and single hosts.

        Any other purposes where the routing and addressing
        architecture might play a crucial role, including QoS,
        better security and spam reduction.

        Improving IPv4 address utilization.

     IPv4's future, including use of NAT and map-encap.

     IPv6 adoption, transition mechanisms and future development.

     Transition to IPv6 or to whatever other architecture the
     RRG might recommend.

     Future architectures potentially quite different from IPv4/6
     - provided they involve scalable multihoming, TE etc.

     General discussion of the above and related fields which
     do not fit within the scope of the RRG list, as determined
     by the Moderators.

     This includes discussion of low-level details of potential
     solutions, including nascent ones such as GSE for IPv6 -
     and any speculative theories and proposals.

  Alternative viewpoints to those compatible with the rough RRG
  consensus are welcome and list members may discuss solutions to
  the routing, addressing, future Internet architecture questions
  which differ in goals, principles, details and time-scales from
  those which would fall within the developing RRG consensus.

  Since the discussions may be diverse and hard to keep up with,
  participants are encouraged to post occasional "Summary: xxx"
  messages pointing out what has transpired and linking to the most
  significant messages in the archives.

  Only the Moderators may cross-post messages from one list to the
  other - such as perhaps the occasional "Summary: xxx" to the RRG
  list.


   - Robin



--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg