You are here
"DEODIS" TEAM: Design and Evaluation Of DIstributed Systems
(Shared systems and networks, cooperation, scheduling)
Our DEODIS team works in the design, optimization and evaluation of shared systems. The team addresses subject matter that originates and is applied in multicomponent environments; their approaches are thus oriented towards support or service to applications at the library or middleware level. The targeted platforms cover a broad spectrum, from slightly coupled platforms to network sensors and collaborative systems, to heavily coupled platforms such as parallel systems or clusters of computing grids.
The complexity of the studied systems generally does not permit a formal approach to solving problems. The DEODIS team thus puts its expertise to work proposing conceptual, architectural or algorithmic solutions and evaluating their pertinence and efficiency.
Goals and Research Areas
The team makes use of diverse tools in modeling (ontologies, modeling of problems of scheduling, multiagent systems), algorithm design, (optimal, heuristic) or protocols (safety, energy performance), and optimization (scheduling, parallelism support), oriented towards distributed systems and their specific qualities. Solutions are then evaluated by simulation or through full-scale execution.
The main research topics are:
› Wireless sensor networks and limited-resource systems : safety, optimization of energy consumption and protcols;
› Distributed algorithmics for synchronous collaborative applications : fault-tolerant consensus algorithms, traceability;
› Scheduling and optimization : pipe-line type applications, energy consumption and efficient multiagent systems.
The DEODIS team proposes original solutions, well adapted to the highly specific problem areas of distributed systems. Their entire body of work is broadly aimed at the context to which their solutions are applied and take into account the precise traits of each problem area.
The team thus has a firm grounding in the technologies used in contexts of real application :
› Sensor networks
› Multiagent systems
› Computing clusters