摘要
针对转移法和睡眠法分别存在增加整体服务成本和资源发现速度较慢的缺点,提出了一个基于觉察上下文计算的降低服务成本的方法.该方法先抽象出资源发现与交互协议运行中的上下文,如系统平稳度、时间段、上下文历史,然后确定上下文的获取方法并分析上下文的获取代价,最后根据推理上下文得到的信息有目的地选择合适的方式来降低服务成本,如停止主动广播、睡眠.基于此思想,构建了包含传感器层、中间层、应用层的系统框架.实验结果表明,在最坏情况下,该方法没有增加整体服务成本,较睡眠法而言,它能让“穷”设备多睡眠37 32%的时间,发现资源的速度快15 68%.
To overcome shortcomings of transshipment-method and hibernation-method that increased holistic cost of service and slow the speed of resource discovery, a method reducing cost of service based on context-aware computing was introduced. This method abstracted some contexts in resource discovery and interaction running time, such as system stableness degree, time span, history of context, then, methods of capturing contexts were established and the cost of capturing contexts was analyzed, lastly, effective manners were purposefully selected to reduce cost of service according to information offered by reasoning contexts, such as stopping broadcast and hibernating. A system framework including sensors layer, middle layer, and applications layer was built based on the idea above. The results of experiments indicate that the method does not increase holistic cost of service under the worst condition, allows the poor devices to hibernate longer than hibernation-method by 37.32%, and the speed of discovering resources is faster than hibernation-method by 15.68%.
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2005年第2期154-157,共4页
Journal of Xi'an Jiaotong University