Design of a dual-computer cluster system and availability evaluation

Considering high-availability of the storage system has become a highly desirable property for network attached storage systems, we develop a special software application supporting the dual-computer fault-tolerant scheme. In this paper, we describe the design and implementation of a dual-computer fault-tolerant system. It includes two fundamental techniques: data mirroring and failover. The data mirroring technique is based on logging and tracking scheme. The failover is realized by heartbeat detection and automatically take over functions. According as availability definition, the theoretical analysis for the dual-computer system has been presented. The finds show that the integration of data mirroring and failover techniques can provide higher availability for the network based storage systems. Network managers can turn to this dual-computer scheme as the solution to resolve the increased demand on the file system availability at file server level.