On Mon, Oct 11, 2010 at 12:35 PM, Dave CROCKER <dhc2(_at_)dcrocker(_dot_)net>
wrote:
On 10/11/2010 8:25 AM, Joel M. Halpern wrote:
Without getting into the question of whether your suggestion would have
helped
anything in terms of transition and interoperability, it shares one major
flaw
with the path we did adopt.
There is no incentive to spend resources to get there.
Indeed, it has been remarkable how poor the sales pitch has been to
resource-poor operations that are expected to adopt this, even after all
this time.
[other good stuff I mostly agree with deleted]
The problem is not merely marketing in the sense of messaging. The problem
with each one of the stalled IETF infrastructure upgrades is deployment
deadlock.
Specifically there is a cycle of ungranted requests. Alice has no incentive
to upgrade her infrastructure because she cannot use any new feature until
Bob upgrades. Meanwhile Bob has no incentive to upgrade ahead of Alice.
Mere exhortations from the great and the good have very limited effect.
Specifically, with the IPv6 upgrade we need to catalog each of the
stakeholders that have to upgrade their gear and work out a unilateral
deployment incentive for each one in turn.
Whenever I try to propose this sort of thing to conferences the referees
always kick the papers back saying that it is 'obvious'. Well if it is so
damn obvious, how come we don't seem to be able to manage to do it?
I think it is a fairly obvious consequence of a rational choice model.
Assume that each actor is not going to budge unless there is a short term
benefit to themselves for changing behavior. Now I agree that there are
major differences between the way the world really works and the axioms of
rat choice modeling. But they certainly work as models in cases where there
is a strong positive feedback loop (network effect).
What I want as a consumer is a box that enables me to do things like peer to
peer video chat efficiently and without all my traffic going in and out of
some peer to peer network whose real function is no more than NAT bypass.
Trying to tie functionality to the flavor of network I am on is a losing
proposition for me. I only have one broadband provider in my area at the
moment and they know it.
--
Website: http://hallambaker.com/
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf