Service-Oriented Computing in Sensor Networks

Jie Liu, Feng Zhao

Abstract: Consider the following scenario in an urban business district. A traffic engineer needs to gather vehicle statistics at major intersections for possible improvement to road layout. Police officers want to check out suspicious vehicles as soon as they enter the area. Environmental protection workers monitor smog level and must warn people when the public health is at risk. For economic reasons, these users may all tap into a common pool of sensing sources — a network of networks that connect to different sensors such as vehicle sensors, cameras, and smog detectors scattered across the city. The large-scale networked sensor system is likely to be deployed by multiple venders over time, so that the cost of construction and maintenance may be shared by the different user groups. Such a system may consist of both mobile and stationary nodes with a wide of range of processing and communication capabilities. It must be integrated seamlessly into the Internet and must provide intuitive interfaces for remote user interactions. There will be multiple, concurrent users exercising different functionalities of the system for different purposes. The system will have to be self-monitoring and resource-aware, and has a certain level of autonomy to decide on the best use of available resources to fulfill multiple users’ uncoordinated requests.