Ad-hoc networks hold great promise, but there is a vast array of competing proposals for organizing such networks. Unfortunately it is unclear, in general, how to choose among the various proposed designs in any given deployment. In response, this project is asking a fundamental question: how should a collection of nodes decide what basic architecture to adopt in organizing into a network? Candidate basic architectures include those that organize via traditional routing protocols, employ epidemic protocols, or use delay-tolerant networking protocols. This project approaches the question as one of deciding what state should be maintained, and how long that state should be maintained, in the network. Relevant state includes both control state (properties of the network) and data state (the data in transit). The project is connecting dynamic properties of the network with the nature of control and data state to be maintained. A key project output is an identification of the network and traffic characteristics that can be used to determine an appropriate level of control and data state maintenance. The effect of these characteristics on the different classes of communication strategies is being rigorously evaluated. Using the characteristics that are identified, the project is developing signaling mechanisms to ensure that communication strategies optimally adapt to changing network conditions.
The result of the project will be more robust and broadly effective methods for creating ad hoc networks in new environments. Such environments can include battlefields as well as post-disaster deployments, where networks must be created quickly under unanticipated conditions.