Software Engineering Thesis Proposal
- Gates Hillman Centers
- THOMAS J. GLAZIER
- Software Engineering Ph.D. Student
- Institute for Software Research
- Carnegie Mellon University
Meta-Management of Autonomic Systems
To meet the demands of high availability and optimal performance in dynamic environments, modern systems deploy autonomic or self-adaptation mechanisms. However, increasingly today's enterprise systems are compositions of many subsystems, each an adaptive system. Each subsystem has its own defined objectives, reasoning methods, and adaptation tactics. Additionally, they are often built by different vendors, hosted on multiple platforms, and have different implementations.
Commonly, human administrators handle situations in which the collection of autonomic systems is behaving sub-optimally. To do so, human administrators evaluate the current state of the individual autonomic systems and the environments to determine if the global quality objectives are likely to be met. When the administrator concludes that intervention is appropriate, they generate a plan of changes to the configuration of the individual autonomic systems with the goal of improving the collection's performance against the global quality objectives. However, generating a plan to change the configurations of the constituent autonomic managers is a complex and challenging task.
In this thesis, I develop an automated approach that addresses this ad-hoc process by providing a formal basis for reasoning about changes to the configurations of constituent autonomic systems through the use of a meta-manager. A meta-manager is a higher level autonomic system that provides assurance about and improves the performance of a collection of autonomic systems in common situations currently handled by human administrators on a time scale appropriate for the context. To simplify the implementation of such meta-managers, and make them practical to maintain, I will develop a domain specific language, specialized for the needs of meta-management, and a reusable software framework that implements a specialized MAPE-K closed control loop that exploits the commonality of autonomic sub-systems associated with meta-management.
David Garlan (Chair)
Betty Cheng (Michigan State University)