Design And Implementation Of Cloud Database Performance Testing Tool For E-commerce Applications

With the increase of the complexity of the e-commerce system and the increase of data volume,there is an urgent need to help users choose the most suitable database products to support the efficient operation of the system by testing.The existing testing tools are insufficient for testing requirements such as e-commerce applications,various transaction scenarios,multiple cloud database types,generating a large amount of workload and fast forwarding,etc.Therefore,it is of great significance to design and implement a cloud database performance testing tool for e-commerce applications.By analyzing the characteristics of e-commerce application,the requirements of ecommerce application,various transaction scenarios and many cloud database,a definition of test model is given from four aspects: data structure,data operation,data constraint and evaluation index.According to the database table structure design of different tested objects,a cloud database performance testing tool scheme for e-commerce application is formed.The test tool is composed of six modules: test task generation module,test data generation and import module,workload generation module,monitoring module,test results sorting module and cloud database forwarding interface.The test task generation module is used to analyze test scripts and generate test tasks for different types of cloud databases.During the test task data loading phase,the test data generation and import module generates the test data set according to the influence factors and database type,and imports it into the tested cloud database.During the running phase of the test task,the workload generation module starts a specified number of threads according to the test task.Each thread randomly obtains the executed transaction by parsing the workload file,and judges the database and access type to generate the transaction load.The cloud database forwarding API forwards the request to perform the test task to the specified cloud database under test.The monitoring module implements the function of the real-time monitoring and feedback of the node status of the cloud database and the load of the test end and each node,which provides a basis for performance analysis.The test results collation module collates the test results,through data extraction and data visualization or other methods,to generate a complete test report.The implementation of the testing tools has been verified and analyzed by experiments.Under the condition of verifying the function of the test tool,we tested the performance of My SQL and Mongo DB database clusters.The results show that the test tool can reflect the performance difference of different databases in the two scenarios of searching for goods and searching and buying goods,and provide reference for the selection of database.