Scalable self-stabilization

This paper presents a methodology for a synchronous non-reactive distributed system on a tree topology to stabilize from a k-faulty configuration in a time independent of the size n of the system. In the proposed methodology, processes first measure and compare the sizes of the faulty regions, and then use this information to schedule actions in such a way that the size of the faulty regions progressively shrink, until they completely disappear. We demonstrate that when k processes fail, the stabilization time is OðkÞ. Apart from its applicability to a wide class of problems, the proposed method achieves scalability with a low space complexity of OðD:ðD:k þ log2 nÞÞ per process, where D is the maximum degree of a node. # 2002 Elsevier Science (USA)