Abstract: |
A dynamic change to architecture is an active area of research within the software architecture community. The objective of all architecture reconfiguration, adaptation and evolution is the improvement of the quality attributes of the software architecture. Only recently the reconfiguration of software architecture at runtime has grown up considerably for the construction of reliable evolutionary systems. The structure of these systems is dynamic and continuously changing.
Consequently, architectures must have the ability to react to events and perform architectural changes autonomously. In this thesis, we provide a new approach based on software multi-agent platform. Such agents are used to supervise the architecture, gather information from it and its environment, capture dynamic changes, and manage them. They monitor the components dynamically and adapt them to structural changes in the architecture. They evaluate dynamically the quality attributes of the architecture. This evaluation will ensure the correctness, robustness, security, availability, etc. as the changes take place so that the system conforms to its architecture and remains in conformance throughout its lifetime.
The autonomous intervention of the platform is based on the scenarios developed by the stakeholders. For these purpose, we are also developed a tool which will be used i) to specify the quality scenarios, ii) to implement these scenarios in the platform.
The prototype, “the autonomous platform based on multi-agent technology for reconfiguration management and the associated quality tool for specifying the scenarios”, proves the feasibility of our innovating approach in the field of: auto-reconfiguration, auto-evolution and self-adaptation of software architectures to satisfy the dynamic constraints of the environment of an application and to satisfy the users’ requirements in term of the quality of the services.
|
