2015 | OriginalPaper | Buchkapitel
Energy-Aware GPU-RAID Scheduling for Reducing Energy Consumption in Cloud Storage Systems
verfasst von : Mehdi Pirahandeh, Deok-Hwan Kim
Erschienen in: Computer Science and its Applications
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Recently, big-data processing applications require a fast and energy-aware cloud storage system. Erasure coding in redundant array of inexpensive disks (RAID) consumes a lot of energy and reduces the performance of cloud storage systems. This is because traditional RAID erasure coding performance is limited to sequential nature of CPU. This paper proposes an energy-aware GPU based flash array cloud storage system with higher input/output performance and less energy consumption. The proposed cloud storage system provides GPU–redundant array of inexpensive disks (GPU-RAID) at a target server based on all–flash array storage. GPU-RAID differs from existing RAID in that it reduces the number of CPU cycles and coding time to generate parity data. Experimental results show that the write performance and read performance GPU-RAID is 33% and 41% higher than that of Linux-RAID, respectively. The results also show that the energy consumption of GPU-RAID is 43% less than that of Linux-RAID.