Types and purposes of performance testing

(1) Load Testing (Load testing)

(The goal is to discover the current performance bottleneck of the system)

It can often be defined as a performance test to determine the performance capabilities of a software system by simulating different workloads. The method adopted is to verify the change of the performance index of the system by increasing the system pressure constantly, and to know the " performance inflection point", that is, a performance indicator does not meet the predetermined threshold value of the target.

This method can find the processing limit of the system, can provide the basis for performance tuning, and also can provide reference for determining the same capacity. This kind of method is usually based on or simulation of the system's real operating environment, and the typical business scenario chosen is the same or close to the actual situation.

(2) pressure test (Stress testing)

(To verify the reliability of the system in extreme cases)

Can be defined as a performance test to obtain the maximum service level activity that the system can provide by simulating the limit load. Usually this kind of test is based on the premise that the system is already in a saturated state, such as CPU utilization has reached more than 85%, memory utilization has exceeded 70% and so on. Under this premise, continue to increase the load on the system, observe the performance of the system, whether there is a denial of service, system downtime and other serious anomalies.

This kind of test generally assists in the system fault localization, inspects the system stability and so on.

(3) concurrency test (Concurrent testing)

(Usually the performance test is the concurrency test here.) The goal is to discover the data structures, algorithms, or problems in the database scripts in the code.

Concurrency testing is usually divided into generalized concurrency and narrow-sense concurrency. The generalized concurrency can be understood as multiple users operating different functions at the same time, while narrow-sense concurrency can be understood as multiple users simultaneously operate the same function. Typically used to verify the system's ability to handle deadlocks and resource contention . As a result, load and stress tests are usually carried out simultaneously. Of course, there is also a separate concurrency test for a particular feature.

(4) fatigue Test (Fatigue testing)

(The purpose is to verify the stability of the system for a long time, for example, to test the system for memory leaks, etc.) fatigue testing is generally used to investigate the stability and reliability of the system. Therefore, it can be defined as a performance test to determine the long-running stability of a system in a given environment by simulating a certain number of users who continuously use or access the system for a long period of time. According to experience, it is generally combined with the pressure test, the comprehensive verification system under certain pressure, long-running stability of the situation