An Analytical Study of Opportunistic Lease Renewal
This paper discusses opportunistic renewal , a lease management protocol designed to keep distributed file systems or distributed shared memories consistent in the presence of a network partition or other computer failures. We present an analytical model of the protocol, compare its performance with existing lease protocols, and quantify performance improvements. In addition, this analytical model provides the structure to understand message overhead and availability tradeoffs when selecting lease parameters. We include results demonstrating that opportunistic renewal substantially reduces the network overhead from message costs associated with lease renewal. As a corollary, opportunistic renewal can reduce the lease length at any given network overhead; e.g., by a factor of 50 at 1% network overhead. Lower overhead makes leasing less intrusive and shorter lease periods allow a system to recover from failure more quickly.
Dr. Darrell Long is Professor of Computer Science at the University of California, Santa Cruz. He also serves as Assocaiate Dean for Computing Infrastructure in the Jack Baskin School of Engineering.
He received his B.S. degree in Computer Science from San Diego State University in 1984, and his M.S. and Ph.D. degrees in Computer Science and Engineering from the University of California, San Diego in 1986 and 1988 respectively.
His research interests include distributed systems, particularly high speed storage systems, fault tolerance, performance evaluation and mobile computing. He directs the Concurrent Systems Laboratory which is part of the Jack Baskin School of Engineering. His research is supported by the Office of Naval Research, the Naval Research Laboratory, the National Science Foundation, the Usenix Association and by International Business Machines Corporation (IBM).
He is a member of the Association for Computing Machinery, the Usenix Association, and a Senior Member of the IEEE Computer Society, which until recently he served as Chair of the Technical Committee on Operating Systems.