Multiscale Traffic Processing Techniques for Network Inference and Control
Abstract
This final report overviews Rice University's accomplishments in the project Multiscale Traffic Processing Techniques for Network Inference and Control in the Defense Advanced Research Projects Agency's, DARPA, Network Modeling & Simulation program. The overall goal of this project was the development of new traffic models, analysis techniques, and model-based processing algorithms for dealing with the type of bursty traffic that can saturate a link or network. Rice's new models are based on the powerful theory of multifractals and represent a step forward in traffic modeling technology. In addition to providing a better understanding of the traffic mechanisms that cause burstiness, they have integrated model fitting, synthesis, inference and prediction within one rigorous statistical framework. Their new models, designed to match salient traffic characteristics at many levels of abstraction and over a broad range of time scales, offer realism while remaining analytically and empirically tractable, statistically robust, and computationally efficient.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2006
- Accession Number
- ADA444211
Entities
People
- Edward Knightly
- Richard G. Baraniuk
- Robert D. Nowak
- Rudolf Riedi
Organizations
- Rice University