关键词: 分布式系统, 心跳检测, 多因素, 心跳频率

Abstract: Heartbeat detecting in the distributed system is one of the key technologies of node fault detection mechanism, and the rationality of the heartbeat frequency setting will affect the accuracy and integrity of the fault detection. In big data environment, fault in distributed system is affected by network, node and operation. A comprehensive evaluation model is put forward to improve the rationality of heartbeat frequency setting under the influence of multi-factors. Under this model, the influence of network, CPU occupancy rate and job size is considered simultaneously. On this basis, a self-adaptive heartbeat detecting algorithm based on multi factor evaluation model is proposed. The algorithm can adaptively change the heartbeat frequency according to the network environment, CPU occupation rate and the job size. The optimal scheme of heartbeat frequency is given by combining various factors. Finally, the effect of multi factor on the heart frequency adaptive adjustment is verified by experiments.

Key words: distributed systems, heartbeat detecting, multi factor, heartbeat frequency