Research On Sensor-Grid Prototype And Application In Water Resource Information System

In discussing the methods of applying wireless sensor technology and grid technology in data collecting and processing in water resources information system, this paper tries to improve modernized level of hydrological data processing. WSN (Wireless Sensor Network) seamlessly connected the digital world and the real physical environment, becoming the emergent computing platform. The core of WSN is an independent sensor node or sensor device with small volume, low cost and power consumption, having the ability of sensing, data processing and wireless communication. Since these devices are deployed in extensive fields, WSN gathered abundant data collecting and processing ability. WSN with a large amount of nodes possesses substantial data obtaining and processing ability, thus becoming important distributed computation resources shared by different users and services. In the meantime, grid computing has developed into a standard means of solving large-scale distributed and heterogeneous resources sharing in dynamic virtual organizations. Current grid computing items mostly concentrate in construction of computing grids and data grids. Computation grids provide distributed computing resources to satisfy customers’computing demands, while data grids provide mass accesses for distributed data and storage resources.The emerging domain of sensor grids extends the grid computing paradigm to the sharing of sensor resources in wireless sensor networks. Sensor grids originate from integrating WSN system with traditional grid system. The system should be equipped with the following important characteristics in order to become sensor grids:First, mass data collected by the sensor can be processed and analyzed by computing resources of traditional grids and data storage resources.Second, different users and application programs can effectively share sensor resources, while each user can access a subset of sensor resources during a specified period of time, operate specific application program and collect sensor data of the required type.Third, it can enable sensor devices of embedded processors to have stronger computing power, and effectively download specific tasks like images and signal processing into sensor devices.Finally, sensor grids can provide various resources to seamlessly access grids through universal means. Adoption of advanced technologies like AIT (artificial intelligence technology), data fusion, data mining and distributed databases provides better means to analyze data and strengthen one’s cognition of environment, the result of which can be reversely utilized for optimizing operation of sensors to better react to environmental situations. Therefore, sensor grids are adaptability expansion of grid computation application.Firstly, The paper discussed problems and challenges that the fusion of WSN and grid system, then propose a sensor grid architecture, called the Scalable Proxy-based aRchltecture for seNsor Grid (SPRING), to address these design issues.The key idea is to utilize agent system as interface between WSN and grids. The SPRING framework integrated WSN and grids, possessing flexible architecture, which is not limited by its own characteristics or application demands of specific goals. As it adopts proxy system as the interface between WSN and grids system, the architecture of SPRING can support various sensor devices, even if those are sensors with limited computation capacity. Besides, the architecture of SPRING is extensible, enabling it to integrate various heterogeneous WSN to enter into the grids system,Secondly, to better study various problems in designing sensor grids and improve architecture design of SPRING, we develop a sensor grid experimental platform and construct a prototype system in a projected co-funded by Singapore-MIT Coalition (SMA) and SUN Company. It is composed of a set of sensor nodes, a 54-node Sun schooling based on AMD Opteron processor, and some proxy servers and user systems based on Linux. At the present stage, the prototype system of sensor grids has already successfully completed stages of allocation, coding and testing, and is capable of executing sensor tasks, collecting and processing sensor data.Finally, we developed wireless sensor grid simulation platform based on GridSim and TOSSIM. Through the platform, users can simulate sensor grid resources, resources management in simulation sensor grids and dispatcher algorithms. After handing in service applications of sensor application to the grids, users can obtain grid simulation results and simulation results of WSN application.