The World Wide Web has proven useful in disseminating information. However, it is extremely difficult to track its spread in such a complex network. One simplification is to model it as a small Local Area Network. Previous work has shown that a gossip-based multicast can be an effective tool for modeling the flow of information on a local area network. We examine the effect of different networks within this mathematical framework. We propose a general mathematical model that incorporates the susceptibility and infectivity of a given machine in a local area network. We also use the concept of proportionate mixing to model the spread of information among heterogeneous patches, where each patch consists of users and non-users of an information router.

This is a joint work with Anthony Billups and Wilbert Fernandez.