Continuous Infrastructure Assessment for Key Business Functions in Changing Environments

Abstract

Communications networks are increasingly important to Critical Infrastructure systems including intermodal transportation and bulk electrical power. An evolving environmental and adversarial landscape motivate the need for stakeholders to continually rank network assets. Such assets observe and respond to physical events over time. Our approach adapts work by Bockholt and Zweig to evaluate assumptions underlying graph-theoretic measures of centrality against properties of realworld communications networks. Our data-processing pipeline integrates heterogeneous data sources, represents communications network topologies at each layer of the network protocol stack, and applies network measures relative to communications observed within a given time interval. Results based on packet captures collected from the DARPA RADICS exercises demonstrate improved asset rankings for measures of centrality better aligned with the properties of real-world network protocols.