Research of Software Reliability Test Based on Test Model

Concept of Software Reliability Test

Reliability is an important software quality characteristic. According to the definition given by IEEE, the software reliability is defined as the probability that software does not cause system failure under the specified conditions and in prescribed time. Software reliability test is a kind of test that carried out to ensure, verify and evaluate the reliability of the software in the environment that closes to the real Situation. It has the following characteristics:

• •

  The test environment is high-demanded: similar to the real one when the software is in real use.

• •

  The starting point of designing the software reliability test cases is to find the defect which has a great influence on the reliability.

• •

  The test focuses on the functions that users often used, so test cases are designed according to the probability distribution of users’ using ways.

• •

  The defects found in reliability test can be used to evaluate reliability of the software.

• •

  The requirements of environment and input coverage are very high, which means that the number of software reliability test cases is also relatively large.

The software reliability test is different from the general software function test, which is mainly manifested in the following aspects:

• •

  The type of test is different. The function test mainly uses the black box test method, while the software reliability test type belongs to the statistical test in the black box test.

• •

  The purpose of the test is different. The consistency test inspection focus on real output and the expected output results, checking the software design to meet the needs of the requirements; reliability test is to ensure the software reliability requirements, the starting point of which should be the case design trying to find out software errors that effect the reliability.

• •

  The sampling strategy of the test instance is different. Functional test check whether all the functions can meet the demand or not; the reliability of the products depends on the usage of it, in order to get a more accurate estimation of the reliability of software, the use cases must be in accordance with the actual use of the probability distribution.

• •

  The coverage rate of the test is different. When executing software reliability test, it is generally required that the input coverage is greater than that of ordinary software function test. It shows that software reliability test is a large number test-cases-needed. Therefore, the amount of data that is based on sex testing is far more than that of general functional test.

• •

  The characteristics of the test cases are different. Every use case of reliability test is a sequence of operations, which reflects a path of software running. In functional test, test cases are generated according to requirements, and are tested for specific inputs, describing a sequence of operation sequences.