A mutual exclusion algorithm for a distributed network of autonomous nodes

Mutual exclusion is treated in distributed autonomous environments. The most important property of the autonomous network treated here is its membership variability, that is, frequency occurrence of entries of new nodes and exits of old nodes. Thus, when the network is large-scale, it is not possible for each node to keep up the information of all other nodes. The authors have previously designed a mutual exclusion algorithm of the single shared resource case for distributed environments of autonomous nodes, based on the Chandy-Misra protocol for Dining Philosopher (diners) problems (K. M. Chandy and J. Misra, 1984). Mutual exclusion of multiple shared resource cases is considered. The condition that no more than k nodes can use the shared resource is given. Since the acyclic graph model used in the model can be represented by a marked graph model, this model is used in analysis.<<ETX>>