The field of performance tuning mainly corresponds to system performance tuning. Generally speaking, performance tuning is intertwined with other performance testing application areas; performance tuning has many adjustable objects, and can be tuned at all stages of system application.
For system applications that have been deployed in actual production environments, performance tuning may first focus on adjustments to the application system deployment environment , such as adjustments to servers, database parameters, and application server parameters;
For applications under development, performance tuning will pay more attention to factors such as the implementation method of application logic, algorithms involved in the application, and the design of the database access layer. At this time, it does not require whether it is a test or production environment, as long as the entire tuning It is sufficient to have a test baseline environment for comparison during the optimization process.
A standard performance tuning process is roughly as follows:
1. Determine the baseline environment, baseline load and baseline performance indicators
The benchmark load refers to a standard that can be used to measure and compare performance tuning test results (the environment must be strictly consistent each time a performance test is performed), a running environment, a test operation script, and a standard that can be used to measure tuning effects. Performance.
2. Adjust the system operating environment and implementation methods, and perform tests
This is the core step in the performance tuning process. The purpose is to improve the performance of the application system through adjustments; it mainly includes the following three aspects:
①Adjustment of hardware environment
Mainly adjust the hardware environment in which the system runs, including changing the server and host device environment on which the system runs (switching to a machine with higher performance, or adjusting the total physical memory, number of CPUs, etc. of some servers), adjusting the network environment ( replace
Faster network equipment, or the use of higher bandwidth networking technology), etc.;
②Adjustment of system settings
Mainly adjust the basic platform settings for system operation, such as: adjusting the core parameters of the Unix system according to application needs, adjusting the memory pool size of the database, adjusting the memory size used by the application server or adopting a higher version of the JVM environment, etc.;
③Adjustment of application level
Mainly adjustments to the application itself, including selecting a new architecture, adopting new data access methods, or modifying the implementation of business logic, etc.;
PS: During the actual performance tuning process, specific adjustments will depend on the specific situation; and do not adjust too many parameters or application implementation methods at once, otherwise it will be difficult to judge which adjustment has an impact on performance (generally 3 to 5 are the most good).
3. Record test results and analyze them
This step and the previous step constitute a performance tuning loop, and the exit of the loop is "reaching the expected performance tuning goal".
The main testing methods used in performance tuning are: configuration testing, load testing, stress testing and failure recovery testing.