[e2e] traffic dispersion and blocking probality

[David P. Reed]

At 01:32 PM 4/1/2002 -0500, Naidu, Venkata wrote:
>
>   Almost all of
>   the datalink/physical layers are connection oriented
>   in that sense and there do exists some loss probability
>   as time varying function.

I realize that this is really just an April Fool's Day discussion....  but 
in the spirit of the day, I'll take it seriously.

Yeah, we really want to reintroduce all that circuit-switched stuff because 
it was so important to Internet applications like the WWW, email, IM, etc.

I'm being blunt because somebody has to.

Just because the old telephony textbooks are still in print doesn't mean 
that the idea of formulating communications as circuits with long hold 
times and stable data rates was useful.

All distributions driven by applications are either Poisson or Gaussian, too.