作者: Susan V. Vrbsky , Michael Galloway , Robert Carr , Rahul Nori , David Grubic
DOI: 10.1016/J.FUTURE.2012.12.016
关键词:
摘要: Regardless of whether data is stored in a cluster, grid, or cloud, management being recognized as significant bottleneck. Computing elements can be located far away from the storage elements. The energy efficiency centers storing this one biggest issues intensive computing. In order to address such issues, we are designing and analyzing series efficient aware strategies involving replication CPU scheduling. paper, present new strategy for replication, called Queued Least-Frequently-Used (QLFU), study its performance determine if it an strategy. We also benefits using scheduling strategy, backfilling, which uses job preemption maximize usage allows longer periods suspension time save energy. measure QLFU existing replica on small green cluster running power consumption with without backfilling. Results have demonstrated that reduce negatively impacting response time.