Shulou(Shulou.com)06/01 Report--

In short, the purpose of performance testing is to obtain the response time, throughput, stability, capacity and other information of the system to be tested. I think the discovery of some specific performance-related defects (such as memory overflows, concurrent processing, etc.) is just an additional result. At a higher level, what the performance test wants to find most is the bottleneck. How to get the information you need needs to be tested in many ways.

When we get a performance test task, we first need to be clear about our test goals and output; a start without a plan is doomed to failure. Performance testing requires a standardized process to figure out what needs to be done at each stage, so that we don't get in a hurry when we get a task.

The performance testing process is roughly divided into the following phases:

1. Test preparation

2. Script development and debugging

3. Test execution

4. Evaluation of test results

5. Tracking after testing

Test preparation includes:

1. Test objectives:

Obtain from requirements documents or users: number of concurrent users, transaction throughput requirements, response time requirements, system resource requirements, scalability requirements, etc.

2. Application software:

Obtain from developers: overall system architecture (BS/CS), using development language, communication protocol, dynamic session, CAPTCHA, etc.

3. System deployment platform:

Obtain from operation and maintenance personnel: physical deployment (LAN / virtual machine), hardware architecture (CPU/MEM/IO), operating system, database (version / startup parameters / storage layout), middleware (parameter configuration), deployment mode, etc.

4. System load model:

Through research: select the key use case throughput and user behavior habits from the business level, determine the test process and process business ratio; understand the peak / normal scenario throughput from the system load

5. Make test plan and test plan.

Test script is the basis of performance testing, and monitoring results are the key to performance testing; solid code foundation and skillful collection and analysis of monitoring results are the guarantee of performance testing; after the system is online, it is also a good way to learn and improve by comparing the actual load and performance and the experience in the testing process.

In practice, we often test two types of software: bs and cs. What do we need for these two performance indicators?

The general metrics that Bs structural programs generally focus on are as follows (simple):

Web server metrics:

* Avg Rps: average number of responses per second = total request time / seconds

* Avg time to last byte per terstion (mstes): the average number of iterations per second for business playbooks. Some people will confuse the two.

* Successful Rounds: successful request

* Failed Rounds: failed request

* Successful Hits: number of successful clicks

* Failed Hits: number of clicks that failed

* Hits Per Second: clicks per second

* Successful Hits Per Second: successful clicks per second

* Failed Hits Per Second: number of failed clicks per second

* Attempted Connections: number of attempted links

CS structural programs, because the general software background is usually a database, so we pay more attention to the database test indicators:

* User 0 Connections: the number of user connections, that is, the number of database connections

* Number of deadlocks: database deadlock

* Butter Cache hit: database Cache hit condition

Of course, in practice, we will also look at memory, CPU, and system resource calls under multi-user testing. These indicators are actually one of the extended performance tests: competitive testing. What is competitive testing? software competes to use a variety of resources (data records, memory, etc.) to see its ability to compete with other related systems for resources.

We know that software architecture restricts the choice of testing strategies and tools in actual testing. How to choose the performance testing strategy is what we need to know in the actual work.

General software can be divided into several types according to system architecture:

C/s

Client/Server client / server architecture

Three-tier architecture based on client / server

Distributed architecture based on client / server

B/s

Three-tier architecture based on browser / Web server

Three-tier architecture based on middleware application server

Multi-tier architecture based on Web server and middleware

Summary:

Generally speaking, I think that the difficulty of performance testing is not mainly in the technical means, the Internet era technology is shared, we should be good at searching and using the achievements of others. Even if you can't handle it, there must be professional development engineers, database administrators, and system administrators on the team who can help you with it. The real difficulty is that you have to figure out how to test is effective and guaranteed, which is the most important ability of test engineers.

From the same point of view, thought is fundamental.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.