Pseudo-Random System Testing: Coverage Estimation and Enhancement

摘要

In system level testing, a number of primitives, such as machine instructions, are used as basis for test cases applied to the system under test (SUT). Different combinations of the valid primitives are applied to the SUT as stimulus. The goal is to check whether the SUT executes each instruction correctly. However, the time complexity associated with measuring test coverage proportionally increases with the complexity of the system under test. As a result substantial test time may be spent on collecting and analyzing test coverage related data of a large system. Furthermore, introducing a new test coverage measuring and/or enhancing mechanism into an existing pseudo-random test generation tool can be both time-consuming and difficult. A simple and novel method to estimate and enhance the test coverage of pseudo-random test case generation systems is provided.