以克服现有服务质量度量方法的主观性、反映服务调用中存在的不确定性和各影响因素之间存在的内在关系为出发点,定义了原子服务调用率、成功率和效率3个因素,通过对原子服务调用的历史信息进行统计计算得到各因素的量值,提出一种基于不确定性因素叠加的原子服务质量度量方法,以及基于各原子服务质量平均值的高粒度Web服务的质量度量标准,并给出服务优先级的判断方法.性能分析验证了所提出方法的高效性、可扩展性和可行性.
As an important basis of Web services selection, discovery and composition, the quality of Web services is widely studied and has become a subject of intense debate in recent years. It is necessary to conquer the subjectivity of existing methods for measuring quality of Web services, depict the inherent uncertainties implied in service invocations and the mutual relationships among the deterministic factors. Therefore, invocation rate, success rate and efficiency are defined as three representative factors with respect to the quality of elementary Web services. By statistical computation of historical invocations on elementary services, the quantitative values of the above three factors can be obtained. In this paper, according to the basic ideas of evidence combination and evidence superposition, an approach is proposed to measure the quality of elementary Web services based on the superposition of the above uncertain factors. Based on the mean of corresponding qualities of elementary Web services, another approach is proposed for measuring higher-granularity services. Further, a method for determining priorities among Web services is given accordingly. The proposed methods are well consistent with people's intuition. Performance studies verify their efficiency, scalability and feasibility.