Re: draft-ietf-multi6-multihoming-requirements-03

[Masataka Ohta <mohta@necom830.hpcl.titech.ac.jp>]

Sean;

> | Can't we simply admit the fact that none of us have any operational
> | experience to be able to discuss multi6 requirement document
> | to be used later to evaluate proposals and move on without it?
> 
> I would hope that whatever forum evaluates any proposals from
> any body dealing with the IPv6 site multihoming problem, that
> "working code" is weighted at least as heavily as "rough consensus".

Code of end to end multihoming is already running over the mobile
IPv6 network with more than 150 routers, which is one of the largest
IPv6 network in the world.

So?

> "code" is an abstract shorthand which can be read as "recipe"
> or anything else used to describe actions to be undertaken by
> human beings in the process of cooking up a solution.  it does
> not mean actual snippets of C or APL.

It is often the case that a good solution is not to undertake any
actions.

> If you think you have working code that might not survive 
> scrutiny based on the emerging rough consensus over the
> requirements drafts, *now* is the time to suggest changes.

What I have been seeing is not consensus but tiredness on abstract
discussion based on CURRENT operational environment with bloating
routing table.

So, if there is one thing to be added as an requirement:

	The solution MUST somehow change the current operational
	environment, which may make some or most of the requirements
	based on the current operational practice meaningless.

						Masataka Ohta