DRISP: a versatile scheme for distributed fault-tolerant queues

A versatile scheme for implementing fault-tolerant queues in a distributed system is proposed. Based on the combination of two simple concepts, by distributing replications of a queue and conducting intelligent-sequential probe (DRISP), the backbone of this work is unsophisticated, yet powerful. A description is presented of the protocols for handling a replicated and distributed queue, which is different from a replicated file in the sense that FIFO order should be maintained in a queue. The replicated distributed queue supported by these algorithms provides dynamic fault tolerance, high availability, and uniform load balancing with small storage space requirements and low communication cost. It is also adaptable to environment changes. Consistency is guaranteed