Self-Stabilization is the property of an autonomous process to obtain correct behavior no matter what initial state is given. Thus, self-stabilization automatically corrects following arbitrary transient faults that corrupt the state (so long as the program's code is still intact). Moreover, if correct behavior and program assumptions are such that a process could be instantiated on a number of different network topologies, then self-stabilization implies that the system topology and state can be asynchronously changed, without notifying the process of this change, yet the process will eventually self-correct its behavior to the new situation.

Self-Stabilization is related to Autonomic Computing, which entails several "Self-*" attributes: self-management, self-configuration, self-healing, self-optimization, and self-protection. Abstractly, self-stabilization appears related to traditional topics of control theory (stability of control loops); however, the requirement of recovery from arbitrary transient faults includes arbitrary changes to the state of any presumed controller--the controller must be engineered to self-recover above all. Furthermore, the literature of self-stabilization is mostly discrete and uses techniques from distributed computing research. The book Self-Stabilization by Shlomi Dolev surveys the topic. A rather dated bibliography about self-stabilization is this 1992 bibliography.

Meetings, Conferences, and Events on Self-Stabilizing Systems

The next event will be:

SSS 2024

Past Events