> Keith, who does have the right to redesign the architecture
> of the Internet, and under what circumstances?
>
> This is a serious question.
and it deserves a serious answer.
To make fundamental changes to the architecture of the Internet would
affect a great many people with widely varying interests. Such an
effort would therefore need to be done slowly and deliberately, with
broad input, a great deal of care in its management, sound technical
foundation, a design team or teams of *very* talented people to
evaluate multiple approaches according to previously-established
criteria, with iterated review and feedback from a wide variety of
interests. In short, it would need to try to get at least rough
consensus from the whole technical community.
The IPv6 effort tried to approximate this idealized process. It
was a painful struggle. A lot of people weren't happy with the result,
and hardly anyone is happy with all of the result. But at least
there was an attempt to get broad input, consider a variety of
approaches, and to craft a compromise that served the needs of a
wide range of interests. If we had to do another thing like this
we might handle it better based on what we learned last time around.
But most of us hope we won't have to do something else like this
anytime soon.
As for the circumstances, we'd have to have a shared realization
either that the old architecture could no longer serve our needs
(such as we had with IPv4) or that a new architecture would be so
far superior to the old one that it would be worth the pain of
development and transition.
Contrast this with the kind of effort (and there have been a few
things like this in recent years, so I'm not singling anyone out in
particular) that says "we'll just tweak this one thing here to solve
our immediate problem" even though this tweak creates problems for
other interests who aren't represneted in the working group.
Keith
