Test Unit ( the UT ( Unit Test )): a test unit testing is the basic unit of software carried out, such as a module, a procedure like. The aim is to test the correctness of the basic unit of software. The method of testing the main unit with a control flow testing, data flow testing, debugging test, Fenwick testing.
Integration Test (IT ( Integration the Test )): Integration testing is testing software system integration process carried out, its main purpose is to check the software interface between units are correct. Integration testing strategy mainly top-down and bottom-up two kinds.
System Test ( ST ( System the Test )): the entire product seen as a test of the whole system. The aim is to verify that the system meets the definition of requirements specifications, identify areas do not conform with the requirements specification or with contradictions. Many software system testing method, there are functional testing, performance testing, UI testing.
Acceptance Test ( ATP ( Acceptance the Test Procedure )): After completion of system testing software products, carried out before the product release software testing activities. It is the last stage of testing. Its purpose is to verify whether the system user requirements specification (which may include project or product acceptance criteria) the requirements, test trying to find software defects remaining as possible to help further improve the software, and to ensure that a system or software the final product is user acceptance. Including usability testing, compatibility testing, installation testing, documentation (such as user manuals, operating manuals, etc.) testing several aspects.
Regression Testing: Regression testing is the maintenance phase in software testing after the software modifications. Its purpose is to test modifications to the software are correct. Here, the modified correctness has two meanings: First, the changes made to achieve the intended purpose, such as the error has been corrected, able to adapt to the new operating environment and so on; and second, does not affect the validity of other functions of the software.
Performance Test :( dividing load testing, stress testing, stability testing, concurrent test, the test configuration, benchmarks), usually the first three to, substantially the same as other test, for example, it is a complicated stress, and reliability stability is the same meaning.
Performance Test: refers to the performance expectations of the premise of the constant pressure system, the verification system resources within an acceptable range, if the expected performance can be achieved.
Load Test (Load Testing): refers to the system continue to increase the pressure or increase the duration under some pressure until an item or a number of performance indicators system reaches safety threshold value to determine the maximum load the system can bear.
Stress Test (Stress Testing): refers to the case of exceeding the safety load, increasing pressure on the system, by identifying bottlenecks or can not receive a system performance point of the user's request, in order to obtain the maximum level of the system can provide the service.
Stability test (S tability T esting ): refers to the system to be tested under specified conditions, pressure is applied to the system and some operations run for a longer period, thereby detecting system is stable.
Configuration testing (ConfigurationTesting): by adjusting the system hardware / software environment, understand the various effects of different Polygonatum on system performance, in order to find the optimal configuration of the system. Through environmental understand the impact of different factors on system performance, in order to find optimal transfer method. Performance in the 1, 2 performance of the current configuration, the current configuration is able to adapt to future business growth 3, the current configuration of the existence of tunable resources
Concurrent test (Concurrency Testing): concurrent access by simulating a user, to access the same multi-user test application, module, or data, whether the observation system deadlock, the processing system familiar decreased to some other performance problems.
Reliability Test (Reliability Testing): When the system pressure at a certain business, allow the system to continue running for some time, to observe whether the stability requirements of the system. It must be given clear requirements, such as the system can continue to run without failure a few days. (The same definition stability test, not the same name)
Benchmark (Benchmark Testing): Under certain hardware / software and network environment, a number of simulated users run one or more virtual service, the test result table for the reference data, system tuning or evaluation process, by running just like business scenarios and compare the test results to determine whether the tuning system to select the effect or provide the basis for decision making. Case of the test object 1, 2 metrics improve the performance of the test, the test and tune the system to ensure the required performance or service agreements.
Capacity test (V olume T esting ): similar to the load test, the system performance can be seen as a particular performance in a particular environment, i.e. setting limits or boundaries. Object is analyzed in advance by testing a metric limit (e.g., maximum number of concurrent users, the number of database records, etc.) reflect the software application features, the system software without any failure in the limit state or can keep its primary function normal operation. Testing will also determine the maximum load capacity or workload test object in a given time can continue treatment.
Fatigue strength test (Fatigue Strength testing): Under the system stable operation, can support the maximum number of concurrent users or daily operation the number of users, the continued implementation period of the business to determine the system to process the largest transactions performance indicators and resource monitoring indicators comprehensive analysis process workload strength properties. Fatigue strength test can reflect the performance of system problems, such as memory leaks.
Test failover (Failover Testing): After the problem lies focuses on systems based on pre-defined policies can recover back, and the ability to resume normal operation. Is generally carried out with its load balancing system, mainly in order to test system failure, whether or not have a big impact overall current, the impact is within an acceptable range, and the user can continue to use the system.
Application Performance Testing: proficiency testing, capacity planning, performance tuning, defects found
Response time (R & lt esponse T IME ): refers to the response time required from the request, the client requests from the beginning, the end and from the server in response to the presentation time of the page.
Number of concurrent users: refers to the number of online users to interact with the server at the same time. In the actual performance tests, generally more concerned about the number of concurrent users of the service, the formula (1) and (2) is used to estimate the peak number of concurrent users, where C is the average number of concurrent users, Cp is the peak number of concurrent users, n is the number Iogin session, L is the average length of Iogin session, Iogin session is defined as user logs into the system to exit the system time, the user is assumed to produce Iogin session Poisson distribution, T is the length of time examined. A length of time, for example, examine OA system for 8 hours.
The average number of concurrent users C = n L / T (1)
Peak number of concurrent users Cp = C + 3 √C (2)
Assume OA system has 5,000 users, the average number of users accessing the system every day is 800, users use the system within 8 hours of work time, a typical user in one day from the login to the average time to exit the system for four hours, according to the formula ( 1) and (2) calculate the average peak number of concurrent users and the number of users. Wherein, C = 800 * 4/8 = 400, Cp = 400 + 3 * √C = 460.
Online users: refers to the user logged into the system, the number of users login session (session) is not a failure.
TPS (Transaction Per Second): the number of transactions per second or transactions the system can handle, it is an important measure of system capacity.
CTR (Hit Per Second): The client sends requests to the server.
Throughput (Throughout): the number of bytes per unit time returned from the server, can also refer to the number of requests processed per unit time submitted by the client server. It is an important measure of network performance, throughput = throughput / test time, usually, the larger the value of the throughput, the greater the value of the throughput, the greater the value of the throughput, the more the load capacity of the system strong.
Throughput: the number of bytes processed by the server per unit time, the reaction index server processing capacity in bytes / sec (Bytes / s). F = Nvu * R / T, F represents a throughput, Nvu the VU represents the number of (the number of virtual users), R represents the number of bytes for each request within time T vu issued, T is the time used in the test performance. But if the system encounters a performance bottleneck, this formula will no longer apply. The stronger the greater the throughput branch system capacity; with the increase in throughput should increase hits per second, this increase is based on the bandwidth enough to handle all the requests raised by users on the basis of. If you increase the number of clicks and throughput constant or decreases, it means that the server can not perform the requested increase (hits per second), the result is the transaction response time increases.
Resource Usage: refers to the extent resources are suitable for different hardware server system. Including CPU utilization, memory utilization, disk utilization, network and so on. Analysis system resource utilization is based primarily on improving the performance properties and further, during the tuning test configuration, the tuning is determined by comparing the utilization of system resources preferably disposed before and after the modulation effect. = Actual resource usage amount of resources used / total amount of available resources * 100%
Performance counters (Counter): analyzed by a range of indicators, such as: CPU, memory, etc., this series of indicators called the counter, it is a description of the server or operating system performance of some of the data indicators. Add counters to open the main observation system resource usage. Performance counters Performance counters including operating system, database counter, application server counters. Counter plays a key role in "Monitoring and analysis" in the performance testing process, especially in scalability and system analysis to locate performance bottlenecks, Que counter value plays a very important role, it must be noted that, generally, only a single performance counter reflects one aspect of system performance analysis test results of the performance testing process must be analyzed based on a plurality of different counters.
Think time (Think Time): two request time interval. Loadrunner function is used in lr_think_time ()