Re: Strange Attractors in Network Flows

[Tim Bass <[e-mail]>]

Valdis replies:

> Is the 'ethernet capture effect' what you are thinking about?

I am certainly no expert on chaos theory; but chaotic systems
are typically non-linear dynamic systems that can exhibit a
very high degree of sensitively to small changes.  These systems
are complex, similar to IP internetworks, i.e. the weather forcasting.

The 'ethernet effect' as you mention is only one of potentially
numerous variables and 'effects' in large IP internetworks that
can influence the behavior of the system.  Traditional mathematical
techniques look at complex models such as IP internetworks and
'throw out' the nonlinear parts and overly simplify the model
so 'traditional' techniques of modelling can be performed.

However, error analysis in these systems show that small errors
in calculations (or perturbations) products very large changes
in the behavior of the system (known as the Lorenz 'Butterfly'
effect).

The 'ethernet capture effect' discussion you attached isolates
one particular effect that in weakly causal in IP internetworks
and is not central to IP-CHAOS model development and strange
attractors in dissipative systems.

IP-CHAOS theory would look at the seeming random packet arrival
and departure rates (or flows rates) and attempt to look for
a strange attractor that models the overall behavior of a
complex nonlinear system. (or something like that, since the
concept of IP-CHAOS is something that I have introduced
here on big-internet)... hoping to find that someone has
written a paper on it before so we can benefit from their
prior work in the area.

The paper or references that I am seeking would talk directly to
the ideas of chaos theory related to IP internetworking.

Thanks,

Tim