Method for testing performance of database on basis of business model

Abstract

The invention provides a method for testing performance of a database on the basis of a business model. The method comprises the following steps of establishing a business model based on the business; configuring test information corresponding to the business model according to the business model, and storing the well-configured test information into a file to form a test file; testing the performance of a database which is associated with the target business according to the business model and the configuration information acquired by reading the test file; adopting a result of the performance of the database as a test result to be outputted.

Description

technical field [0001] The present application relates to database testing, in particular to a method for database performance testing based on business models. Background technique [0002] With the continuous development of data services, database performance has largely become a reflection of enterprise efficiency. To understand the performance of the database, we can only complete it through some testing software. [0003] At present, there are two problems in the software of database performance testing. One is that it cannot be customized according to specific business. The second is that the performance of comprehensive and complex test data cannot be tested. One of the existing technologies is software that only has a fixed transaction model for testing. It can test some specific business functions, such as order delivery, but some special businesses cannot be tested well. And there is another one that can only be tested by simple database query, addition, etc., a...

AI Technical Summary

Problems solved by technology

To understand the performance of the database, we can only complete it through some testing software.

[0003] At present, there are two problems in the software of database performance testing. One is that it cannot be customized according to specific business.

Second, the performance of comprehensive and complex test data cannot be

tpcc simulates the bank user transfer model, and our application system model is quite different from the tpcc user transfer model. If we use tpcc to test our database, due to the large difference in the model itself, the accuracy of the test results is not enough. Reliability is not high

The results of sysbench itself are simpler, and only simple sql can be configured for testing. The tables and fields of the database cannot be simulated according to the real business, and the relationship between tables cannot be expressed, and the test results vary greatly.

[0005] At present, the way to truly simulate the business is to run the real application online, through the pressure of the real application test data, and to test completely from the business point of view, it needs to use a large number of application servers for testing, and the ratio of online database to application server is 1: About 50, if you want to simulate the real business to test the bottleneck of the specific database, you need to use a lot of test resources, and the flexibility is also limited, and you cannot test the bottleneck of the final database

Images