Date of this Version
Recently energy consumption becomes an ever critical concern for both low-end and high-end storage server and data centers. A majority of existing energy conservation solutions resort to multi-speed disks. However, current server systems are still built with conventional disks.
In this paper, we propose an energy saving policy, eRAID, for conventional disk based RAID-1 systems. eRAID saves energy by spinning down partial or entire mirror disk group with predictable performance degradation. The heart work of eRAID is to develop an accurate dynamic performance control (including disk power management) scheme. To guarantee service quality, the dynamic performance control works for two performance Measures/response time and throughput. In addition, a time-series analysis model (ARMA) is used to fore- cast workload features while queueing network models are adopted to do performance prediction. Experimental results show that eRAID can save up to 32% energy without violating predefined performance degradation constraints.