Dependability assurance in distributed systems

Distributed computing has the potential of providing highly dependable services by having replicas of the server processes on several nodes in the network. This technology is relevant for network internal services (e.g. network management) as well as external services (e.g. e-commerce). Keeping consistency between replicas is demanding and introduces performance penalties. Fault transparency, one of the virtues of distributed computing, also requires management functionality, which is critical for the QoS, both during fault handling and normal operation. The project studies the means to provide trustworthy services by this approach and the quantitative evaluation of the QoS obtained. The work has hereto focused on QoS related management in the Jgroup/ARM platform and experimental evaluation.

Relations to external projects and activities:
The Jgroup/ARM Project, in cooperation with University of Bologna and University of Stavanger.

Project coordinator: Bjarne E. Helvik
Professor: Bjarne E. Helvik