一种车载云计算中自增长失效检测器

An accrual failure detector in vehicular cloud computing

针对车载云计算多样化的失效检测需求,通过对其系统架构进行分析,提出了一种基于层次式的自增长失效检测器HA-FD,能同时满足车载云计算中多个应用不同服务质量的需求。该检测器利用路侧单元对系统中车辆节点进行分组,形成一种层次式的检测方法,而且直接根据心跳消息计算怀疑值,不需任何分布假设。实验结果表明:HA-FD检测器减少了系统中检测关系的数量,降低了检测负载,避免了分布假设对检测准确性的影响,能在保证检测准确性的基础上,以最低的检测负载获得最快的检测速度。

To meet the diverse requirements of failure detection in vehicular cloud computing,a new Hierarchical Accrual Failure Detector(HA-FD)based on the architecture of vehicular cloud computing was proposed.HA-FD can meet the requirements of different service qualities of multiple applications simultaneously.In HA-FD,the nodes are grouped by Roadside Unit to implement a hierarchical detection method,and the suspicion level is calculated by direct handling of the heartbeat without any distribution assumption.Experimental results show that the HA-FD decreases the detection relationship of the detection system,so as to reduce the overhead.It avoids the influence of distribution assumption on the detection accuracy.It also improves the detection time by using minimum overhead.