Title :
Metrics, Models and Methodologies for Energy-Proportional Computing
Author_Institution :
Dept. of Comput. Sci., Virginia Tech, Blacksburg, VA, USA
Abstract :
Improving non-peak power efficiency has the potential to significantly enhance the efficiency of a data center and allows us to host more resources under a given power budget. In this paper, we use RAPL interfaces to analyze and model the performance (both throughput and response time) of SPECweb benchmark under subsystem-level power limits. We show that performance under a subsystem-level power limits can be modeled using simple and well-studied non-linear models. We then leverage a load prediction model and an optimization framework to create a runtime system for power management of enterprise application. Our work shows that effective subsystem-level power capping improves the energy proportionality of the server.
Keywords :
benchmark testing; business data processing; computer centres; power aware computing; RAPL interfaces; SPECweb benchmark; data center; energy-proportional computing; enterprise application; load prediction model; nonlinear models; nonpeak power efficiency improvement; optimization framework; performance analysis; power budget; power management; response time; running average power limit interfaces; runtime system; subsystem-level power capping; subsystem-level power limits; throughput; Benchmark testing; Load modeling; Measurement; Optimization; Random access memory; Runtime; Time factors; Energy-proportional Computing; Enterprise Workloads; Green Computing; RAPL;
Conference_Titel :
Cluster, Cloud and Grid Computing (CCGrid), 2014 14th IEEE/ACM International Symposium on
Conference_Location :
Chicago, IL
DOI :
10.1109/CCGrid.2014.119