Granularity-based Uncertain QoS Partitioning for Web Service Reliability

Abstract

The performance of Web services may fluctuate due to their invocations in dynamic environment. Thus, quality of service (QoS) is inherently uncertain when reflecting non-functional features of Web services. QoS has been considered as a significant criterion for selecting those functionally similar services. In this paper, we proposed a novel granularity-based partitioning model (GPM) that is applied to evaluate the reliability of Web services with the consideration of uncertain QoS via cloud model and coefficient of variance. Extensive experiments with 1,558,224 service invocation records have been conducted to validate the effectiveness of our proposed approach. The results demonstrate that it outperforms the state-of-the-art traditional ones.