Research and Improvement on Failure Detection Algorithm
Xinjia Zhang, Lan Luan, Lu Han, Zhong Lu · 2008
Pervasive computing has to be resilient to various kinds of faults and should be able to function despite faults. However, fault tolerance issues have not been well explored so far in pervasive computing research. In this paper, the traditional failure detection algorithm based on heartbeat and some improved failure detection algorithms based on fresh point are discussed and compared. Further, this paper constructs a failure detection algorithm based on fresh point by second layer mode, and gives an improvement by using the historical misjudgment probability to adjust the second detection time dynamically. Finally, experiment is demonstrated about the three algorithms and the results show that the improved algorithm can provide higher accuracy of failure detection in distributed systems.