Active Virtual Network Management Prediction: Complexity as a Framework
for Prediction, Optimization, and Assurance
IEEE Computer Society Press, Proceedings of the 2002 DARPA Active
Networks Conference and Exposition (DANCE 2002), page 534--553. San Francisco, CA, (May 2002)ISBN 0-7695-1564-9.Abstract
Research into active networking has provided the incentive to re-visit
what has traditionally been classified as distinct properties and
characteristics of information transfer such as protocol versus
service; at a more fundamental level this paper considers the blending
of computation and communication by means of complexity. The specific
service examined in this paper is network self-prediction enabled
by Active Virtual Network Management Prediction. Computation/ communication
is analyzed via Kolmogorov Complexity. The result is a mechanism
to understand and improve the performance of active networking and
Active Virtual Network Management Prediction in particular. The
Active Virtual Network Management Prediction mechanism allows information,
in various states of algorithmic and static form, to be transported
in the service of prediction for network management. The results
are generally applicable to algorithmic transmission of information.
Kolmogorov Complexity is used and experimentally validated as a
theory describing the relationship among algorithmic compression,
complexity, and prediction accuracy within an active network. Finally,
the paper concludes with a complexity-based framework for Information
Assurance that attempts to take a holistic view of vulnerability
analysis.
