Autonomic resource provisioning in rocks clusters using Eucalyptus cloud computing
Ekasit Kijsipongse, Sornthep Vannarat · 2010
Data centers always face challenges of peak and fluctuating resource demand from time to time. Building a data center that is large enough to meet a peak demand is not cost effective. The emerging of Cloud computing model allows the data center to dynamically acquire additional resources on demand and pay only for what resources having been used. So, the data center has a flexibility to satisfy users with higher QoS while keeping the investment on IT infrastructure at an affordable cost. In this work, we present an architecture of autonomic provisioning in High Throughput Cluster (HTC) computing system using resources from Cloud. The system is developed as an extension to Rocks Clusters. It allows a data center to transparently and securely extend a local cluster into remote Cloud computing resources on demand through dynamic provisioning mechanism. We also introduce a set of provisioning policies that are aware of the different resource requirement in each job and adapt the system accordingly. Experiments have been carried out in our testbed to show that the proposed system is self-configured and self-organized.