A Distributed Locking Protocol

Abstract

Concurrent accesses to a file frequently occur in a distributed computing environment where a few number of network-attached servers are designated as a data storage pool and the clients are physically connected to the servers via network, like GigaEthernet or Fibre Channel. In such a distributed computing environment, one of the major issues affecting in achieving substantial I/O performance and scalability is to build an efficient locking protocol. We present a distributed locking protocol that enables multiple client nodes to simultaneously write their data to distinct data portions of a file, while providing the consistent view of client cached data, and conclude with an evaluation of the performance of our locking protocol on a Linux cluster. © Springer-Verlag 2004.