VMFS: 一种持久性内存统一管理系统
VMFS: A Unified Persistent Memory Management System
- 电子学报 2021年49卷第12期 页码：2299-2306
- 作者机构：
1.南开大学计算机学院，天津 300350
2.天津市网络与数据安全技术重点实验室，天津 300350
- 作者简介：
[ "张佳辰 男，1994年生于河北唐山.南开大学计算机学院硕士研究生.主要研究方向为存储虚拟化和持久性内存. E‑mail:jczhang@nbjl.nankai.edu.cn" ]
[ "胡泽瑞 男，1996年生于安徽池州.南开大学计算机学院硕士研究生.主要研究方向为大数据存储和分布式系统. E‑mail:huzr@nbjl.nankai.edu.cn" ]
[ "赵 盛 男，1998年生于安徽芜湖.南开大学计算机学院硕士研究生.主要研究方向为存储虚拟化和持久性内存. E‑mail:zhaos@nbjl.nankai.edu.cn" ]
[ "施文杰 男，1994年生于山西大同.南开大学计算机学院硕士研究生.主要研究方向为大数据技术. E‑mail:shiwj@nbjl.nankai.edu.cn" ]
[ "王 刚（通信作者） 男，1974年生于天津.南开大学计算机学院教授，博士生导师.主要研究方向为海量信息存储和并行计算. E‑mail:wgzwp@nbjl.nankai.edu.cn" ]
[ "刘晓光 男，1974年生于河北安国.南开大学计算机学院教授，博士生导师.主要研究方向为搜索引擎和区块链系统." ]
- 基金信息：国家自然科学基金项目(U1833114;61872201);天津市科技发展计划(18ZXZNGX00140;18ZXZNGX00200;20JCZDJC00610)
DOI：10.12263/DZXB.20201239
中图分类号： TP316收稿：2020-11-02，
修回：2021-05-06，
纸质出版：2021-12-25
- 稿件说明：
移动端阅览
张佳辰,胡泽瑞,赵盛等.VMFS: 一种持久性内存统一管理系统[J].电子学报,2021,49(12):2299-2306.
ZHANG Jia-chen,HU Ze-rui,ZHAO Sheng,et al.VMFS: A Unified Persistent Memory Management System[J].ACTA ELECTRONICA SINICA,2021,49(12):2299-2306.
张佳辰,胡泽瑞,赵盛等.VMFS: 一种持久性内存统一管理系统[J].电子学报,2021,49(12):2299-2306. DOI： 10.12263/DZXB.20201239.
ZHANG Jia-chen,HU Ze-rui,ZHAO Sheng,et al.VMFS: A Unified Persistent Memory Management System[J].ACTA ELECTRONICA SINICA,2021,49(12):2299-2306. DOI： 10.12263/DZXB.20201239.
摘要
针对当前持久性内存（PM，Persistent Memory）资源管理方案无法兼顾持久化特性和可字节寻址特性的问题，提出了一种融合Linux系统内核虚拟内存系统和文件系统的持久性内存统一管理系统VMFS（Virtual Memory File System）.VMFS中的单个PM分区可同时提供内存分配和文件存储服务，并利用内外存统一管理的特性可实现内存到文件的重映射机制，避免了不必要的数据拷贝，提升了文件读写性能，且维持了原生编程接口.实验结果表明，对比内外存分别使用PM的方案，VMFS有效提升了文件读写性能.在两种实际工作负载下，VMFS相对于使用DRAM（Dynamic Random Access Memory）和PM分别作为内存和存储的方案具有成本优势和一定程度的持久化性能提升.
Abstract
Current resources management schemes of persistent memory cannot take advantage of data persistence and byte-addressability of PM(Persistent Memory) at the same time. We propose VMFS(Virtual Memory File System)
a unified management system based on the virtual memory subsystem and a file system of OS kernel. VMFS provides memory allocation and file storage services for applications using a single PM partition. Based on the unified management of virtual memory and file storage
VMFS supports the data re-mapping between memory and files. As the re-mapping mechanism reduces the number of data copying times
file read and write latency performance is improved. The test results show that compared with the system using PM as memory and storage sperately
the proposed method can accelerate file reading and writing while the native system call interface was maintained. Under two realistic workloads
VMFS also shows cost and performance advantages over the scheme using DRAM(Dynamic Random Access Memory) as memory and PM as storage.
关键词
Keywords
references
RAOUX S , BURR G W , BREITWISCH M J , et al . Phase-change random access memory: A scalable technology [J]. IBM Journal of Research and Development , 2008 , 52 ( 4.5 ): 465 - 479 .
XU C , NIU D , MURALIMANOHAR N , et al . Overcoming the challenges of crossbar resistive memory architectures [A]. PATTERSON D A. Proceedings of the 21st International Symposium on High Performance Computer Architecture [C]. Washington, United States : IEEE , 2015 . 476 - 488 .
HUAI Y , et al . Spin-transfer torque MRAM(STT-MRAM): challenges and prospects [J]. AAPPS Bulletin , 2008 , 18 ( 6 ): 33 - 40 .
HANDY J . Understanding the Intel/Micron 3D XPoint memory [A]. Storage Developer Conference [C]. Santa Clara, United States : SNIA , 2015 .
XU J , KIM J , MEMARIPOUR A , et al . Finding and fixing performance pathologies in persistent memory software stacks [A]. BAHAR I. The Twenty-Fourth International Conference on Architectural Support for Programming Languages and Operating Systems [C]. New York, United States : ACM , 2019 . 427 - 439 .
SPELMAN L . Reimagining the Data Center Memory and Storage Hierarchy [EB/OL]. https://newsroom.intel.com/editorials/re-architecting-data-center-memory-storage-hierarchy https://newsroom.intel.com/editorials/re-architecting-data-center-memory-storage-hierarchy , 2021-05-05 .
BJØRLING M . Operating System Support for High-Performance Solid State Drives [D]. Copenhagen, Denmark : IT University of Copenhagen , 2016 .
DULLOOR S R , KUMAR S , KESHAVAMURTHY A , et al . System software for persistent memory [A]. BULTERMANN D. European Conference on Computer Systems [C]. New York, United States : ACM , 2014 . 1 - 15 .
BEAUCHAMP B , et al . NVM Programming Model [EB/OL]. https://www.snia.org/sites/default/files/technical_work/final/NVMProgrammingModel_v1.2.pdf https://www.snia.org/sites/default/files/technical_work/final/NVMProgrammingModel_v1.2.pdf , 2021-05-05 .
DONG M , CHEN H . Soft updates made simple and fast on non-volatile memory [A]. SILVA DD. Proceedings of the 2017 Annual Technical Conference [C]. Santa Clara, United States : USENIX Association , 2017 . 719 - 731 .
XU J , SWANSON S . NOVA: a log-structured file system for hybrid volatile/non-volatile main memories [A]. BROWN AD. Proceedings of the 14th Conference on File and Storage Technologies [C]. Santa Clara, United States : USENIX Association , 2016 . 323 - 338 .
KADEKODI R , LEE S K , KASHYAP S , et al . SplitFS: reducing software overhead in file systems for persistent memory [A]. BRECHT T. Proceedings of the 27th Symposium on Operating Systems Principles [C]. New York, United States : ACM , 2019 . 494 - 508 .
DONG M , BU H , YI J , et al . Performance and protection in the ZoFS user-space NVM file system [A]. BRECHT T. Proceedings of the 27th Symposium on Operating Systems Principles [C]. New York, United States : ACM , 2019 . 478 - 493 .
BHANDARI K , CHAKRABARTI D R , BOEHM H J . Makalu: fast recoverable allocation of non-volatile memory [A]. VISSER E . Proceedings of the 2016 SIGPLAN International Conference on Object-Oriented Programming, Systems, Languages, and Applications[C] . New York, United States : ACM , 2016 . 677 - 694 .
HU Q , REN J , BADAM A , et al . Log-structured non-volatile main memory [A]. SILVA DD. Proceedings of the 2017 Annual Technical Conference [C]. Santa Clara, United States : USENIX Association , 2017 . 703 - 717 .
COBURN J , CAULFIELD A M , AKEL A , et al . NV-Heaps : making persistent objects fast and safe with next-generation, non-volatile memories[A]. GUPTA R . Proceedings of the 16th International Conference on Architectural Support for Programming Languages and Operating Systems[C] . New York, United States : ACM , 2011 . 105 - 118 .
VOLOS H , TACK A J , SWIFT M M . Mnemosyne: lightweight persistent memory [A]. GUPTA R. Proceedings of the 16th International Conference on Architectural Support for Programming Languages and Operating Systems [C]. New York, United States : ACM , 2011 . 91 - 104 .
HWANG T , JUNG J , WON Y . HEAPO: Heap-based persistent object store [J]. ACM Transactions on Storage , 2015 , 11 ( 1 ): 1 - 21 .
GILES E R , DOSHI K , VARMAN P . SoftWrAP: a lightweight framework for transactional support of storage class memory [A]. COLEMAN S. Proceedings of the 31st Symposium on Mass Storage Systems and Technologies [C]. Santa Clara, United States : IEEE , 2015 . 1 - 14 .
RUDOFF A , et al . Perisistent Memory Development Kit [EB/OL]. https://pmem.io https://pmem.io , 2021-05-05 .
LINUX KERNEL ORGANIZATION , INC . DAX–Direct Access for Files [EB/OL]. https://www.kernel.org/doc/Documentation/filesystems/dax.txt https://www.kernel.org/doc/Documentation/filesystems/dax.txt , 2021-05-05 .
CHEN S H , CHEN T Y , CHANG Y H , et al . UnistorFS: a union storage file system design for resource sharing between memory and storage on persistent RAM-based systems [J]. ACM Transactions on Storage , 2018 , 14 ( 1 ): 1 - 22 .
COLLET Y . LZ4: Extremely Fast Compression Algorithm [EB/OL]. https://github.com/lz4/lz4 https://github.com/lz4/lz4 , 2021-05-05 .
APACHE . Kafka 2 . 8 Documentation[EB/OL]. http://kafka.apache.org/ http://kafka.apache.org/ , 2021-05-05 .
APACHE . MapReduce Tutorial [EB/OL]. https://hadoop.apache.org/docs/stable/hadoop-mapreduce-client/hadoop-mapreduce-client-core/MapReduceTutorial.html https://hadoop.apache.org/docs/stable/hadoop-mapreduce-client/hadoop-mapreduce-client-core/MapReduceTutorial.html , 2021-05-05 .
APACHE . Spark: Lightning-Fast Unified Analytics Engine [EB/OL]. http://spark.apache.org http://spark.apache.org , 2021-05-05 .
HASEED JM , XIAOYI L , DHABALESWAR K. P , et al . Cutting the tail: designing high performance message brokers to reduce tail latencies in stream processing [A]. NIKOLOPOULOS D. Proceedings of the 20th International Conference on Cluster Computing [C]. Belfast, United Kingdom : IEEE , 2018 . 223 - 233 .
0
浏览量
17
下载量
0
CSCD
- 网络PDF下载
- Meta-XML下载
- BibTeX下载
- Endnote下载
- RefMan 下载
- NoteExpress下载
- NoteFirst下载
关联资源
相关文章
相关作者
相关机构
- 技术支持由北京北大方正电子有限公司提供 京ICP备09064830号-19
京公网安备11010802024621 - 本系统建议在Chrome、 IE9+ 以上版本浏览器阅读本站内容，360浏览器请切换至极速模式
- Cookies帮助我们提供服务并提供个性化体验。使用本网站，即表示您同意我们使用Cookies