A virtualization based elastic model for high performance computing clusters in a Networked Control System

This paper investigates a flexible elastic computing scheme for a class of two-layered Networked Control Systems (NCSs). A high performance cluster architecture with virtual properties that hide the complexity of distributed physical infrastructures is presented. The architecture is support for cloud computing to dynamically deliver heterogeneous computational environments and partition the cluster capacity, adapting to variable demands in a networked control system. Also, a performance model employing cloud resources for elastic clusters is developed, which plans the capacity of the cluster to meet a performance policy (e.g. deadline) or cost request to complete a given work load. The performance of model has been evaluated in the execution of heuristic computing workloads. Finally, the comparison experimental results have demonstrated that the virtualization based elastic clusters constitute a feasible and high performing computing platform for a networked control system.