Common performance testing points these methods and strategies should be kept in mind

Date：2022-03-25 15:08:27Views：627

Any task must first confirm the goal of the task. If you don't know the goal, the result of any effort may not be the final desired result. The same is true of performance test. In the past, although testers have been serious and responsible for the test work from beginning to end, after the test report is released, everyone always feels that there are deficiencies in the United States and has doubts about the test results, especially in the performance comparison test for different test objects with a long time span, there are more or less the following problems:

1. Inadequate test preparation, unclear test objectives and detailed test plan;

2. Lack of test and technical reserves for test objects;

3. The stability and consistency of the test environment are insufficient;

4. The accuracy and representativeness of test data are insufficient;

5. The test description is not concise;

Next, we will discuss how to solve these problems while analyzing the above problems.

Performance test preparation

This is a link that is often ignored by testers. After receiving the pressure measurement task, based on the consideration of various other factors, testers are often eager to make progress and immediately invest in specific testing work. They are very busy with testing, recording and analysis. It is only half of the work that they find that either the hardware configuration does not meet the requirements, the network environment is not ideal, or even the software version is wrong, which makes it difficult to ride a tiger for the moment, It's all because you're not ready for the test. So how should we prepare for performance testing?

There is a need to investigate and analyze software projects, and so is testing. After getting the test task, our first task is to analyze the test task. Before starting the test, we should at least clarify the following problems:

a) What to test or who to test?

b) What questions are we testing or are we trying to figure out or demonstrate?

c) What factors will affect the test results?

d) What kind of test environment is needed?

e) How should it be tested?

Only after carefully investigating the test requirements and carefully analyzing the test tasks can we clarify the problems of the above system. Only on the premise of very clear test tasks and extremely clear test objectives can we formulate a practical test plan.

Clear test objectives and detailed test plan

On the basis of fully understanding the test requirements, it is of great benefit to formulate the test plan as detailed as possible for the implementation of the test. The formulation of test plan is detailed in most professional test books, so this paper will not repeat it.

Test technology preparation

In the current environment, it is unrealistic to require testers to grasp all the software and hardware knowledge in a short time, but usually testers should pay close attention to the research of test tools and test theory. In the test plan, sufficient learning time should be allocated to the research test objects and test tools. Only when we fully grasp the test tools and fully understand the test objects can we implement the test. Jianli's test on wrong understanding, even if you try again, the result is the opposite, and it is difficult to prove the problem, let alone convince users with such a test report.

Configure test environment

Only on the basis of fully understanding the test objects can we know what kind of configuration each test object needs, and it is possible to configure a relatively fair and reasonable test environment (which is particularly important in performance comparison and pressure measurement).

Considering the influence of other factors, such as network lock, network speed, display resolution, database permission and capacity on the test results. If conditions permit, we'd better configure several groups of different test environments.

Acquisition and processing of test data

In all tests, the collection of test data is more difficult, especially for GIS software. Each software has its file format, and some software has several formats. In this case, we can only convert the data in the third-party format into the format of each tested software. At the same time, the data shall be processed to some extent, such as processing data redundancy, processing display style, etc.

If the data will be updated during the test, be sure to back up the data before operation. In addition, the impact of data format and data volume on the test shall be evaluated, and multiple sets of data shall be prepared if necessary.

Finally, we must check the effectiveness of the test data to avoid the impact of damaged data on the test results.

How to conduct performance testing

The preparatory work in the early stage of the test is complicated. Making good preparations for the test is more than half of the test work. However, to produce a convincing test report, we should also correctly grasp the strength of the test, maintain the consistency of the test and improve the accuracy of the test.

Judging the quality of software depends on the ability of software to solve practical applications. Only under a certain test intensity can we test the consumption rate of various software resources, the speed of software operation and the stability of software. By comparing the ability of each functional module of different software to solve practical problems and the efficiency of software operation under different test intensities, we can judge the strength of each module of different software, and even the advantages and disadvantages of the whole software.

After the performance test, the input of all parameters should follow the unified standard. No matter which link, even a little deviation, should be corrected immediately, and we can't take chances. Special attention should be paid to the impact of the external environment on the test results. If the external environment is inconsistent in the whole test process, such as different network speed and machine memory utilization, it may lead to the difference between the test results and the actual situation.

How to summarize performance tests

The end of testing is actually the analysis and processing of test data. No matter how well we do our testing work, if the end user has a pile of disorderly data, it is also a fly in the ointment.

First of all, we'd better select representative data from all the test data and make statistical charts, and then work with developers to carefully analyze the data, find out the problems existing in the software and draw test conclusions. Most users really need scientific and objective test conclusions.

In conclusion, various software performance tests have different scope, high intensity and different bottom. However, as long as we follow the methods discussed in this paper with a serious, objective and scientific working attitude, we can do the test work well. The above is the introduction of performance testing. If you need this, please consult!