Software Engineering Institute

The survivability of large, complex networks such as the Internet is an increasing concern, but such networks are difficult to analyze because they are topologically complex, highly non-linear in their responses, and inherently unbounded (i.e., no node in the network can have global knowledge). To support survivability research, this report will describe how to develop statistically valid networks for analysis and, as an example of their use, applying them to the simulation of virus propagation. It will illustrate the construction of network topologies with GENeSIS, a program written in the programming language Easel. The report will also summarize ongoing significant work in this area of research and give readers insight into how information or artifacts flow through networks and how networks respond to major disruptions.