ABSTRACT

A Replicated File System for Grid Computing
Jiaying Zhang and Peter Honeyman
Concurrency and Computation: Practice and Experience, 2007; 00:1–7.

To meet the rigorous demands of large-scale data sharing in global collaborations, we present a replication scheme for NFSv4 that supports mutable replication without sacrificing strong consistency guarantees. Experimental evaluation indicates a substantial performance advantage over a single server system. With the introduction of a hierarchical replication control protocol, the overhead of replication is negligible even when applications mostly write and replication servers are widely distributed. Evaluation with the NAS Grid Benchmarks demonstrates that our system provides comparable and often better performance than GridFTP, the de facto standard for Grid data sharing.

Keywords: replication; distributed file system, Grid computing