Log-based Web performance test traffic description method

Abstract

The invention provides a log-based Web performance test traffic description method. The method comprises four steps of log data preprocessing, user behavior model establishment, user type distribution extraction and time distribution feature extraction. According to the method, user behavior features, user type distribution features and strength time distribution features are extracted from background logs to describe Web application traffic, so that the real description for test traffic of Web application performance is realized and the requirement, for test flow reality, of the performance test is satisfied.

Description

technical field [0001] The invention relates to the field of web application performance testing, in particular to a description method of web application performance testing traffic. Background technique [0002] With the development of the Internet, various web applications emerge in an endless stream, which have had a profound impact on people's work and life. The web application system is based on the Internet and is user-oriented, resulting in a huge number of users. Coupled with its typical B / S architecture, it faces great performance challenges. Web application performance testing is an important means to ensure the quality of the web application system. It is to test whether the performance of the web application is consistent with the expected goal under the condition of limited time and space resources, to determine whether the system has performance defects, and to find the performance through the test result data. Bottlenecks, improve the process of optimizing s...

AI Technical Summary

Problems solved by technology

[0005] URL sequence, which uses a series of request lists to describe traffic. The most common application is recording and playback technology, which describes its characteristics by recording user behavior on the browser side. This method has the advantages of being simple and intuitive, but it has poor script reusability and path Incomplete coverage leads to poor test authenticity and low effectiveness

[0006] UML model, which can be used in the description of web application traffic to describe the interaction between users and applications in detail. The premise of using this model is that there are complete UML software design documents and specifications, which have relatively large limitations. Based on the UML model It has the disadvantages of poor versatility and abstraction, and cannot describe web application traffic with complex business types

[0007] Markov model, Markov chain is used in physics, statistics, biology and other fields, and it is more widely used in the computer field. There are many variants for testing the performance of web applications. It is a statistical model, but there are some complex

[0008] Graph model, using the graph model to describe web application traffic, the structure of the graph is diverse and flexible, for complex applications, the changeable functions can be flexibly handled, but there are also difficulties in construction and data acquisition

[0009] The above methods and models all have different degrees of defects and deficiencies, the most important of which are the flow characteristics described by the model itself and the lack of model data acquisition, resulting in poor traffic authenticity. The key to the description of web application traffic lies in the selection of models and Acquisition of model data

Images