Supporting adaptation of decentralized software based on application scenarios

Software systems formed by autonomous software entities are greatly different from traditional software systems and it challenges researchers to find effective methods of supporting the adaptation of software systems. In this paper, an approach based on application scenarios is put forward for facilitating dynamic adaptations of decentralized software systems in unpredicted situations. Scenarios offer behavior norms to regulate the behavior of autonomous software entities under specific situations so that software entities can take fitted and coordinative actions when they are confronted with diverse and even unpredicted situations. At the end of this paper, a simulation traffic system is developed and studied. The experimental results show that the adaptability of the system is improved remarkably after application scenarios are deployed. In the case study, the efficiency and scalability of the scenario-based adaptation mechanism are also experimented and analyzed.

► Autonomous components are used to represent autonomous software elements. ► An application scenario-based adaptation mechanism is built for software systems. ► The methods of specifying and implementing application scenarios are described. ► A measurement for the efficiencies of adaptation mechanisms is proposed. ► The efficiency and scalability of the adaptation mechanism are analyzed.