一种WebService响应时间的动态预测方法

Web Service Response Time Dynamic Prediction Approach

随着Web Service技术的迅速发展,具有相同或相似功能的服务在互联网上大量出现.如何预测服务质量,进而支持基于服务质量的服务选择,成为了需要解决的关键问题.其中,响应时间是服务消费者最能直接感受到的服务质量体现,提出一种Web服务响应时间的动态预测方法.该方法根据响应时间的历史数据,基于时间序列分析进行动态预测.实验结果表明,这种方法可以显著提高预测的准确性,更好地支持服务的选择.

With the vigorous development of Web Service technology,services with the same or similar functions have emerged greatly on the Internet.How to predict the quality of service to support service selection has become the key issue.Response time is the most direct customer perception for quality of service.This paper proposes a Web Service response time dynamic prediction approach which based on time series analysis of the historical data.Experimental results demonstrate that this approach can significantly improve the accuracy of response time prediction and better support web service selection.