[rrg] Please respond: Questions from the IESG as to whether a WG forming BOF is necessary for LISP

[David Conrad]

[cc's reduced to keep Noel's head from exploding :-)]

Yakov,

On Jan 22, 2009, at 3:12 PM, Yakov Rekhter wrote:
> The question to ask is whether LISP is an appropriate solution to
> the problems discussed at the IAB's October, 2006 Routing and
> Addressing Workshop.

That is _a_ question, but not necessarily _the_ question.  Perhaps a  
better question is whether or not experimenting with LISP will provide  
insights into appropriate solutions.

> Concerns with scaling and operational properties of these
> techniques have been raised many times before (both at the previous
> BOF, as well as on the RRG mailing list). Yet the LISP proponents
> still did not adequately address these concerns.

Oddly, I'm reminded of http://ars.userfriendly.org/cartoons/? 
id=20090125.

Perhaps the reasons LISP proponents haven't adequately addressed those  
concerns is because they are not addressable with the underlying LISP  
assumptions?  Your criticism feels approximately equivalent to the  
arguments made against IP by telco folks when they'd list all the ways  
in which IP violated circuit switching assumptions.  The appropriate  
response is probably along the lines of "well, duh."  When you change  
underlying assumption, you get a different answer with different costs  
and benefits.  TANSTAAFL.

LISP, as with _any_ loc/id approach I am aware of, requires a some  
form of mapping system to translate locators into identifiers.  You  
can argue about whether you want a pull-based mapping, a push-based  
mapping, or some hybrid, but it is a fundamental assumption that a  
separate mapping system exists.

It can be argued that the requirement for a mapping system will doom  
any loc/id approach (as it was argued that without bandwidth  
reservation from source to destination any packet switched approach  
would be doomed), however I, for one, remain unconvinced and hence  
believe experimentation to find out what the other characteristics of  
loc/id separation might be is worth the effort.

Regards,
-drc