Reed's views, was [e2e] Cannara's views

Jon Crowcroft J.Crowcroft at
Mon Apr 16 08:58:21 PDT 2001

In message <v04220804b700a4b12128@[]>, John Day typed:

 >>And when it really comes down to it, do we really think that our 
 >>brilliant technical insights had anything to do with the success of 
 >>TCP/IP?  Or did it really have more to do with economics and politics 
 >>and timing?  A hard look at the events would seem to indicate that 
 >>things outside the technical had more to do with the current success 
 >>of the Internet than our technical solutions.

good point...

this is something you can't establish a proof for
one way or the other since we
can't re-run history and we can't have two internets, 
but it sure is a reasonable viewpoint.... and itsworth discussing,
since if one wants to move forward from now, it behoves us to
at least try to understand what factors led to success....

another thing to ask is whether some kind of internet technology was
inevitable (i.e. would there have been a massive investment in
something different, e.g. alternative energy sources, or 
hydroponics, or whatever)) had we not got at least some elements of
the internet so right...

actually, another thing to look at is technology stability - how many
technologies (e.g. transport/internal combustion engines, telephone, tv etc)
are actually stable over more than 10/20 years anyhow at the level of
detail we're talking about? and more, over their _first_ 10/20 years?)

what i find depressing, given most of the system is _software_.
we seem unable to figure out how to do graceful upgrade....

so the interesting thing that is happening is that while we play games
with level 2, 3 and 4 , to work around our shortcomings, the internet
is showing how general the idea of "routing around damage" is, and is
devising level 7 mechanisms - the whole overlay approach is quite fun,
(unlike the underlay approaches) in that it can be used to side step
any amount of firewall and lower level addresstranslation munging....

in some senses, the end2end principle is a tautology in that it
describes a property of the internet that is intrinsic to anything
that offers internetworking 

of course, at the level of technology specifics, it would be a lot
nicer if we had GSE addressing deployed and TCP fixed to use it (with
MITs cute secured, migratory end point ideas in place)....but then
we'd have no fun with all the CDN work we can now use to workaroudn
the whole hole we've dug ourselves....

on the economic front, i wonder if the tech. stock downturn means that
we can actualy work on upgrades rather than feature creep?



More information about the end2end-interest mailing list