TOWARDS AUTOMATED REGRESSION ANALYSIS AND MANAGEMENT THROUGHOUT SOFTWARE LIFE-CYCLE
Ankur Kumar Gupta, Veena Tripathi · INTERNATIONAL JOURNAL OF NEXT-GENERATION COMPUTING · 2015
Software application performance tends to degrade over a period of time due to introduction of new features, feature enhancements and bugs-fixes as part of the software maintenance lifecycle. Critical applications cannot afford the performance degradation. However, teams engaged in software maintenance are not too effective in detecting potential performance impacting issues during the test and release phases of the software engineering process, since the run-time environments for individual customers are difficult to simulate. This challenge is exacerbated for applications with a large installed-base. This research paper proposes the concept of Application Baselining as a means to effectively detecting and containing software regression. It provides indication of real-time application performance by monitoring its critical parameters over long periods of time. By keeping track of the changes made to the application and its environment, their impact on application performance is correlated. The changes which adversely impact the application performance are then rolled-back to mitigate their effect. Early work towards development of such a framework is presented.