Self-adaptation of a query service using reconfigurable components

Over the last ten years, evolutions in nomadic terminals and mobile networks led to ubiquitous and pervasive computing. Existing query evaluation and optimization techniques for distributed databases using global schemas and cost models are no longer relevant. Query service, the key element allowing users to access data, must adapt itself to frequent changes in pervasive environments (application constraints, user preferences, connectivity, etc.). In this paper, we propose a flexible architecture of a self-adaptive query service using Fractal reconfigurable components. Our proposition is based on two definitions of adaptability; the first one is a static adaptability by providing five different personalities of a query service, the second one is a dynamic adaptability controlled by a rule based mechanism. The architecture also provides adaptive query evaluation techniques in order to improve performance.