DocumentCode :
2537896
Title :
Adaptive energy-aware design of a multi-bank flash-memory storage system
Author :
Du, Yehua ; Cai, Ming ; Dong, Jinxiang
Author_Institution :
Inst. of Artificial Intelligence, Zhejiang Univ., China
fYear :
2005
fDate :
17-19 Aug. 2005
Firstpage :
311
Lastpage :
316
Abstract :
An architectural solution to reducing energy consumption of flash-memory storage system is to take advantage of a multi-bank architecture instead of a single-bank one. Thus, allowing unused banks to be placed into low-power operating modes can help reduce energy. In this paper, an adaptive energy-aware bank assignment policy is proposed to cluster data with temporal affinity into the same bank, increasing the number of banks put into low-power modes. Then we propose an energy-aware garbage collection policy, which could keep banks in idle conditions for long durations of time to compensate for resynchronization penalties. Experiments over the prototype show the effectiveness of these policies and indicate that large energy savings can be achieved with low overhead.
Keywords :
flash memories; storage management; adaptive energy-aware bank assignment; energy consumption reduction; energy-aware garbage collection; multibank flash-memory storage system; Artificial intelligence; Degradation; Delay; Embedded system; Energy consumption; Energy resolution; Energy storage; Nonvolatile memory; Power system reliability; Prototypes; embedded system; energy-aware; flash-memory;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Embedded and Real-Time Computing Systems and Applications, 2005. Proceedings. 11th IEEE International Conference on
ISSN :
1533-2306
Print_ISBN :
0-7695-2346-3
Type :
conf
DOI :
10.1109/RTCSA.2005.18
Filename :
1541100
Link To Document :
بازگشت