一种保证传感器网络实时服务的中间件机制

传感器网络技术的发展导致了在应用与网络之间建立中间件，根据系统资源及网络能源保证在网络上并发运行的应用的QoS。首先，本文结合传感器网络的特性及实时应用的特点，定义了一个事件驱动的保证实时QoS的传感器网络中间件框架。然后，定义了传感器网络中实时应用的执行模型，提出了一种基于时间戳相关的实现技术，在应用调度中结合应用语义保证实时QoS、数据流的时间一致性及并发应用执行的逻辑一致性。最后，给出了基于时间戳模型的调度算法。通过性能测试结果分析表明，它能够显著提高系统的运行效率及对高速外部采集数据处理的实时性，特别适应于处理间歇、高速的传感器网络外部采集数据。

A Middleware for Providing Efficient Real-time services in Sensor Networks

With the continuing advances in sensor networks and application design, appropriate middleware is needed to provide efficient QoS for coordinate concurrent applications on the sensor networks according the amount of resources and sensor energy. Usually, it is required that an instance of real-time application in sensor networks is periodically activated and must be completed before the next application instance is activated, and this results in data lost while acquiring rate is fast. In the paper, we first propose an event-triggered middleware framework and an executing model of real-time applications in sensor networks. And then, by using the timestamp of collection data to educe the timestamp of coordinate concurrent tasks and integrating application knowledge to tailor the design and implement of the middleware, an implement technique based on timestamp-reliance is presented to providing efficient data services and real-time response while guaranteeing time consistency and correctness of concurrent applications. Finally, we design an effective scheduling algorithms based on the model. Simulation test results show that by using the algorithms, the efficiency of data stream processing and real-time response can be enhanced, and the technique is suitable to process intermittent and high speed acquisition data in sensor networks.