Robustness Testing Methods And Tools For Operating Systems

With the fast development of information technology software products have been applied into many fields in our society. Meanwhile, we have higher requirements for the capability and reliability of software products. So the qualities of software products have become our focus. In some pivotal applications (such as book system in civil aviation, balance system of bank, securities dealing system, control software for automatic aviation, military affairs defense and safe control system of nuclear power plant etc.), the use of bad quality software may bring catastrophic effects.Software testing is an important part in the progress of software development. It can confirm the quality of software. In addition, it can provide information such as the feedback information provided to developers. At present, software developments generally use commercial component off-the-shelf (COTS). The use of COTS benefits improvement of efficiency of development. However, it meets many exceptional cases during using COTS in some environments. So in some abnormal cases, it would lead to failure such as collapse, hanging up etc. Therefore, the improvement of the robustness of software becomes more and more important. The robustness is an important factor impacting the quality of software. As a representative, operation systems are the kernel in computer systems. So, its robustness is very important.It is helpful for constructing the high quality software to precisely understand and improve the robustness of operation systems. Now, there is not any precise and repeatedly utilized method to scale the robustness of system. The robustness benchmark has been brought forward to solve this problem. It consists of a group of test cases. Many robustness testing methods at present are introduced in this paper. The dimensionality model used to analyze the robustness testing results is discussed. We use Linux as an example, a novel robustness testing method called package testing is proposed, and a robustness test platform is developed.