Leveraging Self-Adaptive Dynamic Software Architecture

Abstract

Software systems are growing complex due to the technological innovations and integration of businesses. There is ever increasing need for changes in the software systems. However, incorporating changes is time consuming and costly. Self-adaptation is therefore is the desirable feature of any software that can have ability to adapt to changes without the need for manual reengineering and software update. To state it differently robust, self adaptive dynamic software architecture is the need of the hour. Unfortunately, the existing solutions available for self-adaptation need human intervention and have limitations. The architecture like Rainbow achieved self-adaptation. However, it needs to be improves in terms of quality of service analysis and mining knowledge and reusing it for making well informed decisions in choosing adaptation strategies. In this paper we proposed and implemented Enhanced Self-Adaptive Dynamic Software Architecture (ESADSA) which provides automatic self-adaptation based on the runtime requirements of the system. It decouples self-adaptation from target system with loosely coupled approach while preserves cohesion of the target system. We built a prototype application that runs in distributed environment for proof of concept. The empirical results reveal significance leap forward in improving dynamic self-adaptive software architecture.