Modeling software failures and reliability growth based on pre & post release testing

Abstract

During testing process of any SDLC, efforts are made to successfully find out flaws and make the system bug free, but at times while removing the existing limitations some new bugs might embed into the software system. In order to have a competitive advantage, a firm has to quickly come up with its offering despite the presence of some issues in the software. Therefore, it cannot be assured that the software has been completely made fault free at the time of its release. Of late the concept of patching has started to receive adequate attention and has served the firms as a helping hand to deal with these issues (leftover faults). Software developers are making use of this element as a tool for successful survival in market and as such continue the testing process; post release of their product. However, while providing a patch service, it may happen that while fixing the original problem at hand, some other flaws might get revealed due to the fact that original fault was not removed perfectly. With this frame of mind, in this paper, we put forward a methodical framework to study the concept of fault generation before the release of software and after the release when patch service is being provided. The concept has been empirically tested on real life software failure data set. Further, we have compared our results with existing model present in the literature.