Dynamic Adaptation of Data Distribution Policies in a Shared Data Space System

Increasing demands for interconnectivity, adaptivity and flexibility are leading to distributed component-based systems (DCBS) where components may dynamically join and leave a system at run-time. In this paper we focus on the way that data distribution is handled in such dynamic systems. Current architectures for DCBS provide fixed policies for distributing data between components. Consequently, they cannot cater for changes in usage patterns that may result from changes in the configuration of the components of the system. In previous work, we proposed an architecture for a distributed shared data space that provides a suite of distribution strategies. We showed that simultaneously supporting different strategies for different usage patterns improves overall performance. In this paper, we study the dynamic adaptation of distribution policies based on the evolving behaviour of applications. We experimentally demonstrate the benefits that may be gained by dynamic adaptation of distribu...

• Type: Conference
• Year: 2004
• Where: COOPIS
• Authors: Giovanni Russello, Michel R. V. Chaudron, Maarten van Steen
• Keywords: COOPIS 2004, Distributed Shared Data, Dynamic Adaptation, Information Management, Usage Patterns