Title :
RASSD: A dynamically reconfigurable active storage device for energy efficient data analytics
Author :
Ali, Ahmad ; Saghir, M.A.R. ; Akkary, H. ; Artail, Hassan ; Hajj, Hazem ; Awad, Maher
Abstract :
We study the performance and energy characteristics of a reconfigurable active solid-state drive (RASSD) consisting of a tightly-coupled FPGA/SSD pair. The FPGA implements a compute node that uses partial dynamic reconfiguration to implement hardware accelerators that process data streaming from the SSD. Using K-means clustering as a representative data analytics workload, we show that a RASSD node is 1.3-15.2 times faster than a commodity, multi-core, computer, and consumes 9.4-201.9 times less energy. This suggests that using RASSD nodes in contemporary data centers can have a significant impact on performance and energy efficiency.
Keywords :
client-server systems; computer centres; data analysis; field programmable gate arrays; local area networks; power aware computing; reconfigurable architectures; LAN; RASSD node; compute node; data centers; data streaming process; dynamically reconfigurable active storage device; energy characteristics; energy efficient data analytics; field programmable gate arrays; hardware accelerators; k-means clustering; local area networks; partial-dynamic reconfiguration; performance characteristics; representative data analytics workload; tightly-coupled FPGA-SSD pair; Computer architecture; Computers; Distributed databases; Energy consumption; Field programmable gate arrays; Hardware; Instruction sets;
Conference_Titel :
Energy Aware Computing Systems and Applications (ICEAC), 2013 4th Annual International Conference on
Conference_Location :
Istanbul
DOI :
10.1109/ICEAC.2013.6737642
