[RRG] Elegance and the rejection of SHIM6 host-based multihoming

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

Hi Iljitsch,

There are several messages I want to respond to, and will next week,
but I wanted to first challenge something you wrote in "Re: [RRG]
Renumbering...":

> Stephen Sprunk wrote:
>
>>> The users leaned on the RIRs to do PI, and any attempt to get
>>> rid of PI would take even more ooomph than stopping them, and
>>> there clearly wasn't enough for that.
>>
>> Speaking as one of "them", the reason we proposed (and accepted)
>> IPv6 was that the IETF had provided no acceptable (to us)
>> alternative.
>
> That is nonsense. The shim6 effort was well under way at that
> point but not yet mature enough that it was possible to know
> whether it would solve the problem.


The functional goals of SHIM6 are well known and have been for a
while: to provide host-based multihoming.

End-user networks generally don't want host-based multihoming - they
want network based multihoming so an unknown number of hosts in
their network keep operating without a glitch when one link to an
ISP goes down.  They also need a solution which works for today's
Internet - IPv4.  I guess they require any changes to their
operations not to involve host changes or hosts being involved with
network functions such as routing and forwarding.

Also, people want address space which is portable between ISPs - so
they don't have to renumber their network, DNS etc. when selecting a
new ISP.

I believe these are perfectly reasonable desires and needs.  SHIM6
and IPv6's automated router and host renumbering just doesn't do
what most end-user networks need.


> So even if you are of the opinion that shim6 isn't an acceptable
> solution, at the point the decision was made this was still
> unknowable and thus the decision was wrong.

The limits of the usefulness of SHIM6 and IPv6's automated host
renumbering system are and were well known, so I think it is wrong
to state that people making a decision based on their needs and
SHIM6's limits made the wrong decision.


> (And the fact that this decision is made by random people in one
> part of the world rather than by the IETF or some other world wide
> organization with at least _some_ technical props shows how broken
> the RIR policy development system is.)

I guess that if the IETF was working on some promising techniques
which would have provided whatever it is that people need when they
need portability and network based multihoming - but with PA space -
then perhaps folks would have waited before approving the release of
PI space to end-users.

The IETF had and still has no such plans.  When people say they need
to keep their address space, either when selecting a new ISP and/or
for multihoming, it is a perfectly reasonable thing for a network to
need.   This can't be achieved with PI space.

LISP, APT, Ivip, TRRP and Six/One Router are attempts at providing
end-user networks with what they want and need: portability and
network-based multihoming with PI space, but in a scalable way.


I think it would be an unjustifiable over-reaction to require that
all hosts change their stacks and applications just to solve a
scaling problem in the DFZ, which is what ILNP and I guess some
other proposals involve.  Perhaps some of those proposals would
result in a cleaner, more elegant, Internet structure than those
which involve no host changes.

Most people, including me, are not prepared to change everything to
achieve something which optimised "elegance".  We just want the
Internet to keep working well with minimal disruption.

Elegance in English would involve a radical revamp of the language.

Elegance in driving would involve all countries agreeing to either
drive on the left or on the right.

Elegance in Western music notation would involve changing the crazy
C, C#/Df, D, D#/Ef, E, F, F#/Gf, G, G#/Af, A, A#/Bf, B nomenclature.
  (Backwards compatible with some pieces of music written 400 years
ago . . . )  Likewise the horrible layout of piano and organ
keyboards (also backwards compatible with 17th century organs) - a
different chord and melodic fingering pattern for each of the 12
transpositions - would be replaced by something like a 5 row button
accordion system, where a single chord or melody pattern is used no
matter what the transposition.

Elegance in keyboards would involve replacing the QWERTY keyboard
with something better.

(The driving, piano keyboard and computer keyboard questions are
also major questions of health and safety, with the former problem
causing death and injury and the latter two contributing to RSI.)

I wish all these things would happen, and I wish the Internet had
full end-to-end connectivity in a manner that a simple routing
system could scale indefinitely to support portability, multihoming
and mobility for billions of end-user networks.

However, since there are solutions which can provide scalable
portability, multihoming and mobility without host changes - and
since an elegant redesign would only be useful to most end-users
after the prohibitive hurdle of everyone changing their OSes and and
applications, and connecting to the new network - I don't support
such a radical revision of the Internet.  Even if I thought a
radical revision was a good idea, there's no way we could either
impose it on the world or persuade sufficient people to willingly
adopt it (unless we could somehow make it seamlessly interwork with
the current Internet, current applications etc.).


  - 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