Performance Modelling of a Self-adaptive and Self-optimising Resource Monitoring System for Dynamic Grid Environments

As the number of resources on the Grid tends to be very dynamic and potentially large, it becomes increasingly important to efficiently discover and monitor them. This paper investigates ways in which the scalability and performance of Grid discovery and monitoring services can be enhanced. Benchmarks are collated and the performance analysed at the client of the MDS3 Index Service when different notification (push-based) rates are utilised. Moreover, using autonomous concepts, two self-adaptive algorithms are implemented, based on findings from the performance benchmarks. Different workload models are also used as part of several application scenarios, where the notification rate of data is dynamically modified, based on the overhead costs at the MDS3 Index Service. Experiment results are subsequently shown when varying notification update mechanisms and decision parameters are used.