SCLFD: A Novel Latency Fault Detection Approach for Service Composition
Zhenjie Zhang, Zhanfeng Shen, Xiaobin Xu, Hong He · 2023
Due to the uncertainty of the cloud environment, the response time of service inherently exhibits fluctuations and randomness, making it challenging to ensure the accuracy of latency fault detection of service composition. Therefore, this paper proposes a latency fault detection approach for service composition. Firstly, a global response time monitoring framework is introduced by analyzing the execution process of service composition. Secondly, according to the acquired monitoring data, an inference algorithm based on evidential reasoning(ER) rule for response time anomaly alarm is designed. We evaluate the approach with an actual service composition scenario about elevator traction wire rope design. Experimental results show the accuracy of the proposed approach can achieve 91.57%, significantly outperforming traditional approaches.