On-demand attribute-based service discovery for mobile WSANs

Adequate service discovery is required to support dynamic composition as offered by recent light-weight service platforms for wireless sensor and actuator networks (WSANs). To provide flexible and precise service discovery in such dynamic environments, the operational context of services should be taken into account. This context can be inferred by reusing the information collected by the sensor nodes' environmental and resource monitoring services. In order to be usable, a light-weight abstraction of the contextual parameters is required to provide standardized context descriptions. Furthermore, the service discovery process itself needs to be adapted to handle the flexibility of this approach. In this paper, we present a service discovery mechanism that supports the expression of contextual constraints through the concepts of attributes and dictionaries. We report on the operation of such a service discovery solution and show its feasibility in a proof-of-concept implementation on SunSPOT. In this implementation, we integrate service and route discovery into a single process hereby reducing the required network overhead.