Interimistic data dissemination

Abstract

We present low-overhead protocols for disseminating streaming data in an interimistic environment in which the availability of computers, networks, and other resources is very unpredictable. An important example of such an environment is the state of computing and communication resources immediately following a natural disaster or a similar disruptive event. A key characteristic of such environments is the continual change in network membership and topology. Our protocols are based on using the intensional and extensional overlap among the data requirements of network nodes. We describe our testbed, CoDD, that implements these protocols and provide an experimental evaluation of the techniques.