A reactive architecture for cloud-based system engineering

Abstract

Software system engineering is increasingly practised over globally distributed locations. Such a practise is termed as Global Software Development (GSD). GSD has become a business necessity mainly because of the scarcity of resources, cost, and the need to locate development closer to the customers. GSD is highly dependent on requirements management, but system requirements continuously change. Poorly managed change in requirements affects the overall cost, schedule and quality of GSD projects. It is particularly challenging to manage and trace such changes, and hence we require a rigorous requirement change management (RCM) process. RCM is not trivial in collocated software development; and with the presence of geographical, cultural, social and temporal factors, it makes RCM profoundly difficult for GSD. Existing RCM methods do not take into consideration these issues faced in GSD. Considering the state-of-the-art in RCM, design and analysis of architecture, and cloud accountability, this work contributes: 1. an alternative and novel mechanism for effective information and knowledge-sharing towards RCM and traceability. 2. a novel methodology for the design and analysis of small-to-medium size cloud-based systems, with a particular focus on the trade-off of quality attributes. 3. a dependable framework that facilitates the RCM and traceability method for cloud-based system engineering. 4. a novel methodology for assuring cloud accountability in terms of dependability. 5. a cloud-based framework to facilitate the cloud accountability methodology. The results show a traceable RCM linkage between system engineering processes and stakeholder requirements for cloud-based GSD projects, which is better than existing approaches. Also, the results show an improved dependability assurance of systems interfacing with the unpredictable cloud environment. We reach the conclusion that RCM with a clear focus on traceability, which is then facilitated by a dependable framework, improves the chance of developing a cloud-based GSD project successfully.