Evaluation and Analysis of Web Service Based on Heavy Tailed Distribution
Xun Yuan, Liping Huang, Bin Zhang · Advances in engineering research/Advances in Engineering Research · 2015
More and more web services are composited together to accomplish more complicated tasks.Replaceable web services with same function need to be evaluated effectively, so as to select the best one.This paper selects better service by response time, which is easy to collect on client side machine.According to collected data, this paper points out that response time jumps periodically in a long time period, and in a short time period the response time shows heavy tailed distribution which contains normal distribution.This paper puts forward that if we use tail weight of response time distribution, together with average value of response time, we will evaluate more precisely the web service reflecting uncertainty of web networks.This paper gives method of computing tail weight and comprehensive evaluation metric.Experiments show the validity of the method.The method can also be used for evaluation of web services, web services composition, and cloud service.