A power efficient persistent storage consolidation algorithm for cloud computing

Persistent data storage is a necessity in local cloud computing architectures. This research investigates how these architectures can reduce their energy consumption by dynamically scaling the number of storage nodes needed to accommodate user´s needs. While a user is actively using cloud computational resources, their accommodating persistent data is stored on hot storage nodes that are powered on. Inactive users have their persistent data stored on powered down, cold storage nodes. Design of the proposed cloud architecture considers network throughput and storage capacity of storage nodes a priority. This paper introduces a persistent data consolidation algorithm with decreased power consumption. As the cloud architecture increases in size, the power savings are compounded. Also, since consideration is given to decreasing power consumption, this paper addresses the data latency tradeoff of the proposed consolidation scheme.