Approximate Solution Approach and Performability Evaluation of Large Scale Beowulf Clusters

Beowulf clusters are very popular and deployed worldwide in support of scientific computing, because of the high computational power and performance. However, they also pose several challenges, and yet they need to provide high availability. The practical large-scale Beowulf clusters result in unpredictable, fault-tolerant, often detrimental outcomes. Successful development of high performance in storing and processing huge amounts of data in large-scale clusters necessitates accurate quality of service (QoS) evaluation. This leads to develop as well as design, analytical models to understand and predict of complex system behaviour in order to ensure availability of large-scale systems. Exact modelling of such clusters is not feasible due to the nature of the large scale nodes and the diversity of user requests. An analytical model for QoS of large-scale server farms and solution approaches are necessary. In this paper, analytical modelling of large-scale Beowulf clusters is considered together with availability issues. A generic and flexible approximate solution approach is developed to handle large number of nodes for performability evaluation. The proposed analytical model and the approximate solution approach provide flexibility to evaluate the QoS measurements for such systems. In order to show the efficacy and the accuracy of the proposed approach, the results obtained from the analytical model are validated with the results obtained from the discrete event simulations.