Autostabilisation et protocoles réseau

National audienceIn 1974, E.W. Dijkstra defined self-stabilization as the property for a distributed system to recover by itself a correct behavior in a finite number of steps, starting from any initial state. Thus, self-stabilization is a simple and efficient way to tolerate transient faults or failures. This paper surveys works that propose self-stabilizing solutions to problems arising in Computer Networks, such as routing and transport layers, or network control protocols. We also review techniques to design self-stabilizing protocols, and mechanisms that reduce the stabilization time when the number of hitting faults is small