Towards Automated Monitoring and Forecasting of Probabilistic Quality Properties in Open Source Software (OSS): A Striking Hybrid Approach

Abstract

In this paper, we propose a hybrid approach based on the aspect-orientation methodology and time series analysis to the runtime monitoring and quality forecasting of OSS. Specifically, the major objective of this work is to combine the idea of time series analysis with the area of software quality assurance of OSS in which statistical techniques for analyzing of time series is used to facilitate the prediction and forecasting (the term 'prediction' and 'forecasting' are interchangeably used in the literature) of probabilistic quality properties, which are difficult or inapplicable to be evaluated by current approaches such as testing, and also help to increase the reliability and productivity of working OSS system components (towards trustworthy open source software development) requiring extreme runtime quality control. Furthermore, in order to reduce the human effort and to cope with more sophisticated scenarios, this study also aims to automate the analysis and modeling process by providing appropriate tool.