[e2e] End-End Principle rant

[David P. Reed]

Noel - I like what you wrote a great deal.  I'm putting a reference to it 
on SATN.org (my blog) now that there's a place to refer to.

It captures something I have been trying to point out for a long time that 
resulted from people equating the end-to-end argument with the metaphor of 
a "dumb" (Gilder) or "stupid" (Isenberg) network.   That metaphor is 
limited to the view of the network from the customers' point of view (the 
network doesn't do me "smart" favors), but is wrong when applied to the 
internal view of the network (it's worth devoting lots of computation to 
routing, etc. as long as that computation isn't specialized to a subset of 
the possible applications).