A Framework for Sensor-Federated Networking: SenSORCER

Despite many technology advances, the limited computing power of sensor encumbers them from taking part in Service Oriented Architectures. In recent years, the sensor-networking subject has been extensively studied, with the focus on efficient usage of energy in sensors and efficient usage of sensor node processing power. However, providing sensor nodes with the relevant computing power and sensor composition is seldom addressed. With the availability of the meta-computing environment (SORCER), sensors can participate in service oriented computing, by exporting sensor probe data with relevant processing as a service.

The framework introduced in this thesis allows building highly scalable and dynamic sensor networks, providing sensors with corresponding computing environment. We describe a novel approach using Jini network technology to integrate several different sensors and sensor networks into one virtual service-oriented sensor network. The framework makes use of dynamic provisioning mechanism provided by the Rio framework, which autonomically allocates resources to the sensor service. Furthermore, the dynamically typed language, Groovy, provides the runtime computing mechanism involving variables of sensor services. This framework is generic and extensible and allows incorporating different standard and non-standard sensor technologies. It also helps expedite the development process for service-to-service (S2S) applications involving sensor data.