One Research For UMMs-based Statistical Testing Method Of Web Applications

With the prevalence of network and the progress of IT,Web application becomes a mainstream application model;therefore how to ensure the Web application's reliability becomes increasing important.The Web application is characterized as the following:complex structure,a great quantity of code and enormous interlink Web pages.In the limited time and devotion,the traditional,coverage-based testing is undoubtedly invalid.In fact,different portion of software are executed with a high frequency than others.Statistical testing is just based on this fact and aims to identify and adjusts these portions accordingly,subjecting more frequently executed section of the code to thorough testing.By improving security and reliability of the key section,the security and reliability of the system will be promoted so as to improve the efficiency of testing.The prerequisite of statistical testing is to generate the usage-model that precisely reflects the behavior and usage of the system.In the past,the usage-model was mainly constructed by estimation and predetermination,so the behavior and usage of the system can not be reflected precisely.The paper presents a improved Markov model—Unified Markov Model algorithm.The method is on the research of getting the state usage frequency from log files of the system using FastStats,building the UMMs of the system to generate the precise usage-model and guiding the process of Web statistical testing.Based on the introduction and research of the characteristic of the statistical testing,the characteristic of the Web application and the present testing methods in Web environment,the paper issues the idea of one Statistical Testing Method for UMMs-based of Web Applications—UBTM, in this way we can generate the precise and succinct usage-model and guide the process of Web statistical testing.Using the Web usage and failure information from existing Web logs, the usage information is used to build UMMs.The related failure information is used to measure the reliability of Web applications and the potential effectiveness of statistical Web testing,build a evaluation method based on reliability.At last we applied this approach to analyze one Web application—SRES.This paper describes in detail how to build the UMMs of the system step by step and withdraw the failure information from the failure log files.The related data demonstrated the viability and effectiveness of our approach.Also we compare this research with the statistical testing method based on Markov chain usage models,the result proves the superiority of this research.