Opportunistic application flows in pervasive environments

Summary form only given. Supporting applications in sensor-based pervasive environments requires a programming and management paradigm where the behaviors as well as the interactions of applications elements (sensors, actuators, and services) are dynamic, opportunistic, and context, content and capability aware. We present a programming model that enables opportunistic flows in pervasive environments. The model builds on content-based discovery and routing services and defines associative rendezvous as an abstraction for content-based decoupled interactions. Cascading local behaviors (CLB) then build on associative rendezvous to enable opportunistic application flows to emerge as a result of context and content based local behaviors. We also present the design, prototype implementation and experimental evaluation of the Meteor programming framework and content-based middleware.