利用SSD和DRAM构建的扩展内存架构

doi:10.3969/j.issn.1001-2400.2017.03.025

Abstract

Abstract:

The data-intensive applications of Big Data era are eager to boost system performance through large capacity of memory. However, the use of the large DRAM is restricted by its high power consumption and price per bit. Flash as an existing technology of Non-Volatile Memory, it has some advantages of large capacity, low price and low power consumption. Therefore, we propose an extended memory architecture which expands the RAM with the SSD. The large extended memory is managed at an application object granularity to boost the efficiency of accessing data on the SSD. To reduce the waste of physical memory because of the object granularity, it provides a flexible memory partition and multi-mapping strategy to manage the physical memory by micro-pages. We implement a runtime library to provide several interfaces for applications to access the large memory transparently. Experimental results with several data-intensive workloads show that this method can provide up to 3.3 times performance improvement over the one that uses the SSD as the system swap device which works at the page level.

Key words: big data, data-intensive application, non-volatile memory, solid state disk, memory extension

WANG Liyu;CHEN Lan;HAO Xiaoran;WANG Qiang;NI Mao. Memory-extended architecture based on the SSD and DRAM[J].Journal of Xidian University, 2017, 44(3): 144-150.

References

［1］ SALKHORDEH R, ASADI H. An Operating System Level Data Migration Scheme in Hybrid DRAM-NVM Memory Architecture ［C］//Proceedings of the 2016 Design, Automation and Test in Europe Conference and Exhibition. Piscataway: IEEE, 2016: 936-941.
［2］ SUN Z W, JIA Z P, CAI X J, et al. AIMR: an Adaptive Page Management Policy for Hybrid Memory Architecture with NVM and DRAM ［C］//Proceedings of the 2015 IEEE 17th International Conference on High Performance Computing and Communications. Piscataway: IEEE, 2015: 284-289.
［3］ CHEN F, KOUFATY D A, ZHANG X. Hystor: Making the Best Use of Solid State Drives in High Performance Storage Systems ［C］//Proceedings of the International Conference on Supercomputing. New York: ACM, 2011: 22-32.
［4］ KIM Y, GUPTA A, URGAONKAR B, et al. HybridStore: a Cost-efficient, High-performance Storage System Combining SSDs and HDDs ［C］//Proceedings of the IEEE International Workshop on Modelling, Analysis, and Simulation of Computer and Telecommunication Systems. Piscataway: IEEE, 2011: 227-236.
［5］ SESHADRI V, PEKHIMENKO G, RUWASE O, et al. Page Overlays: an Enhanced Virtual Memory Framework to Enable Fine-grained Memory Management ［C］//Proceedings of the International Symposium on Computer Architecture. Piscataway: IEEE, 2015: 79-91.
［6］ KGIL T, MUDGE T. FlashCache: a NAND Flash Memory File Cache for Low Power Web Servers［C］//Proceedings of the International Conference on Compilers, Architecture and Synthesis for Embedded Systems. New York: ACM, 2006: 103-112.
［7］ SAXENA M, SWIFT M M. Flashvm: Virtual Memory Management on Flash ［C］//Proceedings of the 2010 USENIX Conference on USENIX Annual Technical Conference. Berkeley: USENIX Association, 2010: 14.
［8］ GUO W C, CHEN K, FENG H, et al. MARS: Mobile Application Relaunching Speed-up through Flash-aware Page Swapping ［J］. IEEE Transactions on Computers, 2016, 65(3): 916-928.
［9］ WANG C, VAZHKUDAI S S, MA X, et al. NVMalloc: Exposing an Aggregate SSD Store as a Memory Partition in Extreme-scale Machines ［C］//Proceedings of the 2012 IEEE 26th International Parallel and Distributed Processing Symposium. Piscataway: IEEE, 2012: 957-968.
［10］ BADAM A, PAI V S. SSDAlloc: Hybrid SSD/RAM Memory Management Made Easy ［C］//Proceedings of the 8th USENIX Conference on Networked Systems Design and Implementation. Berkeley: USENIX, 2011: 16.
［11］ BADAM A, PAI V S, NELLANS D W. Better Flash Access via Shape-shifting Virtual Memory Pages ［C］//Proceedings of the 1st ACM SIGOPS Conference on Timely Results in Operating Systems. New York: ACM, 2013: 1-14.
［12］ MEZA J, LUO Y, KHAN S, et al. A Case for Efficient Hardware/Software Cooperative Management of Storage and Memory ［DB/OL］. ［2016-04-10］. repository.cmu.edu/ece/370/.
［13］ KAWATA H, OIKAWA S. A Feasibility Study of Hybrid DRAM and Flash Memory Management Unit ［C］//Proceedings of the 2014 IIAI 3rd International Conference on Advanced Applied Informatics. Piscataway: IEEE, 2014: 694-698.

[1]	CAI Tao,LIU Peiyao,WANG Jie,NIU Dejiao,HE Qingjian,CHEN Zhipeng. Highly concurrent NVM storage system based on the asymmetric lock [J]. Journal of Xidian University, 2020, 47(2): 67-74.
[2]	WANG Xing;HUANG Xiaoyu;LIU Xuanpu;KONG Xianguang;NIU Meng. Multi-layer incremental feature extraction method for industrial big data [J]. Journal of Xidian University, 2018, 45(4): 106-111.
[3]	KONG Xianguang;ZHANG Xiong;MA Hongbo;CHANG Jiantao;NIU Meng. Real time feature extraction method for complex industrial big data [J]. Journal of Xidian University, 2016, 43(5): 70-74+152.

Memory-extended architecture based on the SSD and DRAM

PDF (PC)

Like

Knowledge

Abstract

Cite this article

share this article

References

Related Articles 3

Metrics

Comments

Recommended 10