Workshop III: Beyond Internet MRA: Networks of Networks

November 3 - 7, 2008

Overview

Activities of enterprises typically involve multiple networks: ranging from (1) transportation of energy, materials, and components to (2) power grids, supply chains, and control of transportation assets, and to (3) communication and data networks. The networks’ activities are correlated because they are invoked to support a common task, and the networks are interdependent because the characteristics of one determine the inputs or constraints for another. They are becoming even more correlated and interdependent as they shift more and more of their controls to be information intensive and data network-based.

While this “networks of networks” concept enables enormous efficiency and flexibility (both technical and economical) it also has a dark side — by requiring increasingly complex design processes, it creates vastly increased opportunities for potentially catastrophic failures, to the point where national and international critical infrastructure systems are at risk to large-scale disruptions due to intentional attacks, unintentional (but potentially devastating) side effects, the possibility of (not necessarily deliberate) large cascading events, or their growing dependence on the Internet as a “central nervous system”.

This trend in network evolution poses serious questions about the operation and reliability of these critical infrastructure systems in the absence of an adequate theory. The purpose of this workshop is to bring together domain experts from the fields of engineering, biology, mathematics, and critical infrastructure protection to develop the foundation of a nascent theory in support of the networks of networks concept. In particular, we will use the Internet as a case study to illustrate how early verbal observations and arguments with deep engineering insight have led via an interplay with mathematics and measurements to increasingly formal statements and powerful theoretical developments that can be viewed as of what we envision to ultimately become a full-fledged “theory” for Internet-like systems.

Topics of interest include (but are not limited to):

  • a deeper understanding of the implications that the original design philosophy behind the DARPA Internet protocols have had on today’s Internet and of how that philosophy has constrained it’s evolution,
  • the importance of measurements for reverse-engineering the Internet; i.e., learning about its design, functionality, dynamics, and complex feedback regulation by studying its implementation,
  • the role of measurement-based networking research to provide profound theoretical explanations of empirical phenomena through innovative multi-scale representations of network traffic and network structure that respect the architectural design principles of the Internet,
  • the development of synthesis theories like “layering as optimization decomposition” that allow for forward-engineering of new protocols and architectures.

Organizing Committee

David Alderson (Naval Postgraduate School, Operations Research Department )
John Doyle (California Institute of Technology, Control and Dynamical Systems)
Ramesh Govindan (University of Southern California (USC), Computer Science)
Craig Partridge (BBN Technologies)
Walter Willinger, Chair (AT&T Technologies, Engineering Research Center, Mathematics)