Energy-Efficient Streaming Using Non-volatile Memory

The disk and the DRAM in a typical mobile system consume a significant fraction (up to 30%) of the total system energy. To save on storage energy, the DRAM should be small and the disk should be spun down for long periods of time. We show that this can be achieved for predominantly streaming workloads by connecting the disk to the DRAM via a large non-volatile memory (NVM). We refer to this as the NVM-based architecture (NVMBA); the conventional architecture with only a DRAM and a disk is referred to as DRAMBA. The NVM in the NVMBA acts as a traffic reshaper from the disk to the DRAM. The total system costs are balanced, since the cost increase due to adding the NVM is compensated by the decrease in DRAM cost. We analyze the energy saving of NVMBA, with NAND flash memory serving as NVM, relative to DRAMBA with respect to (1) the streaming demand, (2) the disk form factor, (3) the best-effort provision, and (4) the stream location on the disk. We present a worst-case analysis of the reliability of the disk drive and the flash memory, and show that a small flash capacity is sufficient to operate the system over a year at negligible cost. Disk lifetime is superior to flash, so that is of no concern.

[1]  Luca Benini,et al.  A survey of design techniques for system-level dynamic power management , 2000, IEEE Trans. Very Large Scale Integr. Syst..

[2]  Mahmut T. Kandemir,et al.  Energy-aware data prefetching for multi-speed disks , 2006, CF '06.

[3]  HensgenDebra,et al.  Intel Turbo Memory , 2008 .

[4]  Anand Sivasubramaniam,et al.  Power management of enterprise storage systems , 2005 .

[5]  Michael F. P. O'Boyle,et al.  Automatic performance model construction for the fast software exploration of new hardware designs , 2006, CASES '06.

[6]  H. Rothuizen,et al.  A Vibration Resistant Nanopositioner for Mobile Parallel-Probe Storage Applications , 2007, Journal of Microelectromechanical Systems.

[7]  Frank Bellosa,et al.  Cooperative I / O-- A Novel I / O Semantics for Energy-Aware Applications , 2003 .

[8]  Bruce Jacob,et al.  Memory Systems: Cache, DRAM, Disk , 2007 .

[9]  Jeanna Matthews,et al.  Intel® Turbo Memory: Nonvolatile disk caches in the storage hierarchy of mainstream computer systems , 2008, TOS.

[10]  Michael Isard,et al.  A design for high-performance flash disks , 2007, OPSR.

[11]  Alan Jay Smith,et al.  Software strategies for portable computer energy management , 1998, IEEE Wirel. Commun..

[12]  Luca Benini,et al.  Quantitative comparison of power management algorithms , 2000, Proceedings Design, Automation and Test in Europe Conference and Exhibition 2000 (Cat. No. PR00537).

[13]  Trevor N. Mudge,et al.  FlashCache: a NAND flash memory file cache for low power web servers , 2006, CASES '06.

[14]  Scott A. Brandt,et al.  A Hybrid Disk-Aware Spin-Down Algorithm with I/O Subsystem Support , 2007, 2007 IEEE International Performance, Computing, and Communications Conference.

[15]  David E. Culler,et al.  Proceedings of the 5th Symposium on Operating Systems Design and Implementation , 2022 .