Re: net questions for radio broadcast

Little Robbie Burcham (burchamr@lark.cc.ukans.edu)
Sun, 23 Jul 1995 13:28:15 -0500 (CDT)


On Sun, 23 Jul 1995, Bob Hill wrote:

> My experience in listening to KJHK Kansas has been that the current
> quality is comparable to AM radio in most cities. Lots of people listen
> to AM radio. Unless this usage _actually_ creates such an overload on
> the Internet today that it poses a problem (and at least as much of
> problem as is caused by people downloading nude pictures from Playboy's
> and Penthouse's www sites), I think it should be allowed, even
> encouraged, as one of the more innovative experimental uses of the Internet.

In the case of KJHK, the traffic that we produce is not a burden do to
volume, but the *type* of traffic. The type of data we are talking about
is UDP (not TCP) routed by IP. These types of datagrams are not bounded
by the conventions of TCP at all. TCP datagrams are nice to each other,
some will "get out of the way" for others. UDP does not behave this way;
if a UDP datagram is sent, it GOES, and it rolls right over anything in
its path, thus pre-empting other well-behaved traffic in the process.
This is why real-time data encapsulated in UDP is initially frowned upon.
Compound this problem with high-volume data (a la video) and ignorant
modem users, and you get angry SLIP/PPP providers.

The type of traffic involved with the downloading of the nude pictures you
describe above is actually MORE friendly to the Internet than CU-Seeme
traffic. The bottom line is packet networks do fine for the exchange of
data, but once you insert the real-time condition, they still can serve,
but only at the expense of other well-behaved datagrams. Cell networks
on the other hand...

_______________________________________________________________________________
Rob Burcham - BS, Computer Science
University of Kansas "He's nuts! Somebody grab him!"
Department of Special Education
http://robb.sped.ukans.edu
http://www.ukans.edu/~burcham