About the various MySQL performance test results on the network

About the various performance test posts on the web, I would like to say the following points:

1, in order to make the performance test more objective, the actual, should explain to what scene carries on the test, the query, or modifies, whether contains the primary key, contains several indexes, each difference is what. Because of the different MySQL branches, the reason exists because there is a solution to the point exists, rather than for the cottage and the cottage, and even directly take PG to test the conclusion;

2, the test hardware should have a real representative, a lot of testing with vm,1g,2g memory, n old CPU or the CPU of the notebook to test, this kind of test is no typical significance; the actual production machine is no longer the public cloud, that also have Intel E-Series CPU, Enterprise-class hard disk and so on, Many servers also have a BBU, which can greatly affect the performance of different branches;

3, the server parameter setting, PERCONA,MARIADB has the additional parameter control with respect to the Oracle MySQL, these parameter's opening or not also may affect the performance, moreover may have the very small influence in the production, the exclusively uses the same parameter setting to test, also is unreasonable.

So, in reference to this kind of information, colleagues should pay more attention to the context of the conclusion and various assumptions.

About the various MySQL performance test results on the network