Semantic internetworking of sensor systems

Deployed wireless sensor networks are usually tightly coupled with specific tasks and therefore present unique communication needs. With technological advances that are making easy deployment of wireless sensor networks a reality, mechanisms that allow data exchange between sensor systems deployed at different times and initially for different purposes are necessary. Otherwise, sensors which are co-located and can collaborate with one another to increase capacity and/or performance may be prevented from doing so due to the inability to exchange data with one another. To this end we propose a semantic-based routing architecture that enables internetworking of sensor systems. In this architecture each sensor system can define its own address names that reflect application-level needs, but which follows a specific naming structure. Semantic routers are elected in the network and process packets based on both the address name and the naming structure. Processing is based on rules that can be modified according to changes in application-level requirements or changes in the naming structure. We describe this architecture and show that it enables easy internetworking of sensor systems deployed at different times and for different applications, enables data forwarding to only relevant regions in the network (as opposed to network-wide flooding), supports application-level communication needs, and supports hierarchical routing schemes.