T*GreenHDFS: A Cyber-Physical, Data-Centric Cooling Energy Costs Reduction Approach for Big Data Analytics Cloud

Big Data explosion and surge in large-scale Big Data analytics cloud infrastructure have led to burgeoning energy costs and present a challenge to the existing run-time cooling energy management techniques. T*GreenHDFS, a thermal-aware cloud file system, takes a novel, data-centric approach to reduce cooling energy costs. On the physical-side, T*GreenHDFS is cognizant of the uneven thermal-profile in the data centers due to complex airflow patterns, varying ability of the cooling system to cool different parts of the data center, and run-time load distribution. On the cyber-side, T*GreenHDFS is aware of the differences in the data-semantics of the data placed on the clusters. Based on this knowledge, and coupled with its predictive data models, T*GreenHDFS does proactive, thermal-aware data placement, which implicitly results in thermal-aware computation placement in the Big Data analytics cloud compute model. Evaluation results show up to 59% reduction in the cooling energy costs with T*GreenHDFS.unpublishe