Re: [rrg] Constraints due to the need for widespread voluntary adoption

[Robin Whittle <[e-mail]>]

Hi DY,

There's a long history of theoretical discussion about routing in the
RRG.  Your last two messages were highly theoretical.

Do you have a proposal for solving the routing scalability problem?

If so, does it involve changes to host operating systems and/or
applications?  Would your proposal be at odds with any other of these
constraints which arise from the need for voluntary adoption?

  http://www.firstpr.com.au/ip/ivip/RRG-2009/constraints/

If you have a proposal, please describe it in more concrete terms,
with examples etc.

There are many ways of designing a global network which has no
routing scalability problems.  Our task is to provide a fix for
today's Internet - the IPv4 Internet, and for the IPv6 Internet so
that the same scaling problems won't occur if and when it is widely
used.  We can't force people to adopt anything, though there is some
scope for altering IPv6 before anyone starts using it on a
large-scale for general public use.

This task has been likened to converting a propeller-driven passenger
aircraft into a modern passenger jet - while it is flying.

However this analogy does not depict the constraints imposed by the
need for widespread voluntary adoption - adoption which will
typically be motivated by immediate benefits and not by concern about
scalability.

As with the aircraft analogy, the process of change needs to occur
with the network fully operating - so existing addressing and
protocols to hosts, and many other things, need to function normally
during and after the changes.

Some people seem to think it is good enough to solve IPv6's scaling
problems and then wait for the billion-plus users to migrate from
IPv4.  I don't accept this because no-one has been able to convince
me that this mass adoption and therefore non-reliance on IPv4 will
happen any time in the foreseeable future.

If your proposal doesn't alter host requirements at the operating
system or application level - for all IPv4 hosts operating today -
and if it has a chance of meeting the other constraints due to the
need for voluntary adoption, I will attempt to understand and discuss it.

If you don't have a proposal, perhaps you could critique the most
prominent proposals:

   LISP   http://www.lisp4.net
   APT    http://www.cs.ucla.edu/~meisel/apt-rrg.pdf
   Ivip   http://www.cs.ucla.edu/~meisel/apt-rrg.pdf
   TRRP   http://bill.herrin.us/network/trrp-rrg.pdf

all of which I think could meet the 7 absolute constraints imposed by
voluntary adoption, and would work with both IPv4 and IPv6.

  - Robin