In message <136180874894.20020411124804@nexsi.com>, Alex Zinin writes:
>
> Frank,
>
> [...]
>
> > This brings me to the real point of this note :-)
>
> > To me, going into these issues is _way_ beyond what is
> > practical, given the deployed base. Basically, my opinion
> > is that any new architecture that changes hosts and DNS
> > and all that other "stuff" just won't fly. That is, we
> > (the routing research bunch) can not change DNS, can not
> > change the end hosts, etc.
>
> I though we wanted to be revolutionary here :)
> Seriously, I think the research effort should not
> be anchored by existing baggage.
>
> Alex
Guys,
The reason end2end-interest was such a huge success in the IRTF was
that they didn't revolutionize transport in the Internet. They took
on relevant issues and produced practical solutions (large window
solved high speed TCP problems, SACK solved huge window problems,
offshoots of end2end discussions such as MTU-disc, RED, ECN, newreno
TCP, all took on practical problems and proposed practical solutions).
I don't mean to break up the party you guys are having, but you can
only be somewhat revolutionary and still be relevant. If the
transition posed by a solution is a bigger problem than the problem
you are trying to solve the work is unlikely to be relevant.
Curtis
This archive was generated by hypermail 2b29 : Mon Aug 04 2003 - 04:10:04 EDT