Data management for sensor networks

点击翻译按钮获取中文摘要

Sensor networks, can be broadly defined as a network of sensing devices capable of processing, storing, sending and receiving data.   A typical sensor is now a minuscule device capable of processing several data entities simultaneously and autonomously as well as communicating  wirelessly. Sensor networks are typically deployed to monitor physical phenomena,  e.g., temperature and humidity levels, within a geographical area and over a time window (limited by the  network's lifespan).  A few application domains where the use of sensor networks is suitable are: environmental  monitoring, biological contamination detection, warehouse management, traffic organization and battlefield  surveillance.  Being autonomous, sensor nodes have a limited power supply, hence, a main constraint in such an  environment is the energy cost for operating the network.  Given the cost of sending/receiving data -typically  orders of magnitude larger than processing the same in memory- a key issue is to minimize data  communication among nodes while at the same time taking advantage of potential cooperation among nodes in the network when processing queries.   Overall, the aim of this proposal is to investigate, from a database management perspective, techniques for storing and querying data within a SN while minimizing energy cost.  To accomplish this goal  several alternatives, including new algorithms and data structures, will be designed and evaluated.