BROWSE
Detail
Conference Paper
Workload Adaptive Power Conservation Policy
Cited - time in 
Share 
Share
- Authors
- Hyungjik Lee, Jangho Choi, Joonyoung Jung, Dongoh Kang
- Issue Date
- 2013-12
- Citation
- International Conference on Advances in Information Technology (ICAIT) 2013, pp.78-83
- Publisher
- IERI
- Language
- English
- Type
- Conference Paper
- Abstract
- This paper presents a dynamic power management policy for reconfigurable clusters. The existing power management technologies, DVS and DFS, are not very effective at low workload intensity, because servers use about 60% of power of working state despite they are in idle state. To validate our results and policy, we implement them on realistic 3-tier server cluster architecture and use ACPI and wakeup/sleep scheduling policy in proportion to the input workload. Experimental results using our dynamic power management policy using Apache JMeter workload model show that power consumption is saving up to 27% compare to non-policy method.
- KSP Keywords
- Adaptive power, Cluster architecture, Dynamic power management, Management policy, Power Consumption, Power conservation, Server cluster, Sleep scheduling, Workload model, scheduling policies