Cell Relay Archive

Cell Relay Retreat>List Archive>month:1995-Apr> msg00463



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

Re: self-similar traffic model references need

  • From: vjs@calcite.rhyolite.com (Vernon Schryver)
  • Date: Sun, 30 Apr 1995 03:43:51 -0500, Sat, 29 Apr 1995 21:23:31 GMT

In article <ntwkD7t6sv.7B0@netcom.com> ntwk@netcom.com (Network Synthesis) writes:

>A short summary of this property is that data traffic
>follows the human activity of the users. ...

>To carry this farther, in order to understand the behavior of
>computer networks, you need to understand how and why people use
>those networks.  In order to understand how and why people use
>computer networks, you have to understand people.  Because of this,
>when using self-similar models of data traffic, you should 
>understand and acknowledge that the models are not models of 
>collections of machines but instead are models of people 
>interacting with computers and each other.
> ...


That is a fine jape, combining technobabble and politically correct
"concern" for people, but I fear less technically informed readers might
take it seriously.  The phrase "self-simlar" in the famous "self-similar
network traffic" papers has a technical meaning and nothing more to do
with "human behavior" than it does when used in connection with the
orbits of the planets.


>From "On the Self-Similar Nature of Ethernet Traffic (Extended Version)"
in _IEEE/ACM Transactions on Networking_, vol 2, no.1 Feb 1994, page 1:

  ... presence of of "burstiness" accros an extremely wide range of time
  scales: traffic "spikes" ride on longer-term "ripples," that in turn
  ride on still longer term "swells," etc.  This _self-similar_ or
  _fractal_-like behavor ...


Vernon Schryver    vjs@rhyolite.com