基于有效度函数的传感器事务调度策略

Validity degree function based sensor transaction scheduling schemes

针对数据的绝对一致性概念只能够描述二元数据状态的问题,提出了数据有效度函数的概念。给出了两种基于有效度函数的传感器事务调度策略VM-DS与VM-SN。策略通过合理的设置数据对象的有效时间段长度及事务的执行频率来最大化系统中数据的有效程度。通过实验对两种策略的性能进行了对比分析。结果表明,VM-SN在平均数据有效度、有效度负载比等性能指标上均优于VM-DS,能够更好的满足系统的实时性要求。

The disadvantage of the concept of temporal consistency is that it only describes binary states of real-time data objects. The concept of validity degree function is presented to overcome the problem.Based on the new concept,two sensor transaction scheduling schemes VM-DS and VM-SN are proposed.Both schemes maximize the data validity degree by judiciously setting the lengths of data validity intervals and the execution frequencies of sensor transactions. Experiments are conducted to evaluate the performance of two schemes. The results show VM-SN outperforms in terms of the average data validity degree and the ratio of validity degree to workload,thus it can better satisfy the timeliness requirements of the systems.