Performance analysis of three implementation strategies for distributed lock management

A distributed lock manager (DLM) provides distributed applications with a c onvenient means of synchronising their accesses to shared resources. The au thors present a performance study of three different implementation strateg ies for a DLM considering both the layout of the lock database (centralised or distributed) and the strategy used to assign lock masters (static or dy namic). For each implementation strategy, accurate analytical models of com munication cost, resource utilisation and lock request response time are de veloped. The models highlight bottlenecks in the system and show clearly fo r what mixes of incoming lock request types it is best to use static or dyn amic lock master positioning. The analytical formulae are validated against a detailed event-driven simulation which uses realistic hardware parameter s. This validation reveals a good agreement between analytical and simulati on results, particularly with respect to communication cost, node and CPU u tilisation, system capacity and the response time trend.