Exploiting Temporal and Spatial Constraints on Distributed Shared Objects

The advent of gigabit network technologies has made it possible to combine sets of uni- and multiprocessor workstations into a distributed, massively-parallel computer system. Middleware, such as distributed shared objects (DSO), attempts to improve programmability of such systems, by providing globally accessible 'object' abstractions. Early research on distributed shared object systems concerned protocols to maintain consistency across replicated 'memory' objects. Such systems are well suited to scientific applications but have limited support for multimedia or groupware applications. This paper addresses the state sharing needs of complex distributed applications with (1) high-frequency symmetric data accesses to shared objects, (2) unpredictable and limited locality of data access, (3) dynamically changing sharing behavior, and (4) potential data races. We show that a DSO system that exploits application-level temporal and spatial constraints on shared objects can outperform shared object protocols which do not exploit application-level constraints. We describe the features of our S(emantic) DSO and compare three application specific consistency protocols, developed to run on our system against entry consistency for a sample application having the four properties mentioned above