CNS Core: Medium: Collaborative Research: Persistent memory objects for consistent sharing in Non-Volatile Main Memories

CNS 核心：中：协作研究：用于非易失性主存储器中一致共享的持久内存对象

• 批准号: 1900758
• 负责人: Michael Swift
• 金额: $ 50.97万
• 依托单位: University of Wisconsin-Madison
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2019
• 资助国家: 美国
• 起止时间: 2019-10-01 至 2023-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1900758&HistoricalAwards=false
• 关键词: CNS; Core; Medium; Collaborative; Research

The continual performance growth of data center servers is critical to the nation's economic competitiveness and as a catalyst for progress in scientific endeavors. Two important data center components are the main memory, which has fast access but can only store data temporarily, and storage, which keeps data permanently but suffers from slow access. Recent technology advances have brought in a new non-volatile memory, now commercially available, that can both host permanent data and be accessed quickly. However, to reach its potential, utilizing these new memory technologies requires rethinking of how data should be persistently and efficiently stored. This proposal describes a new abstraction for storing persistent data in non-volatile memory: hyperfiles, which are long lived, provide fast access, and can be quickly attached to and detached from a process address space. Hyperfiles provide naming and permission characteristics similar to (but faster than) files, and speed closer to memory. They are accessed directly through loads/stores to avoid system call overhead. This project also investigates new sharing semantics for hyperfiles, allowing non-cooperating processes to share them simultaneously and safely, while keeping the crash recovery property. Architecture support to accelerate hyperfile sharing will also be designed and evaluated. This project will enable fast access to persistent data by reducing overheads inherent in file access, and will enable more concurrent use of data through strong sharing semantics. Big-data workloads becoming common across all industries will either perform better or require fewer resources. The software and prototyping artifacts produced will enable other researchers to further enhance the design, and to adapt workloads to take advantage of the hyperfile abstraction. The education and outreach components of this plan will train the next generation of programmers on use of persistent memory, and will encourage underrepresented students to join the next generation of programmers. The project will maintain data, code, results, publications, and simulator artifacts for the duration of the project plus five years afterward. Each artifact will be released as it reaches sufficient maturity. They will be posted at: https://sites.google.com/view/hyperfile.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

数据中心服务器性能的持续增长对国家的经济竞争力至关重要，也是科学努力进步的催化剂。两个重要的数据中心组件是主内存和存储器，前者可以快速访问，但只能临时存储数据；后者可以永久保存数据，但访问速度很慢。最近的技术进步带来了一种新的非易失性存储器，现在已经商业化，既可以存储永久数据，又可以快速访问。然而，为了发挥其潜力，利用这些新的内存技术需要重新考虑如何持久有效地存储数据。该建议描述了在非易失性内存中存储持久数据的一种新抽象：超级文件，它是长期存在的，提供快速访问，并且可以快速地附加到进程地址空间或从进程地址空间分离。超级文件提供了与文件相似（但比文件更快）的命名和权限特性，并且速度更接近内存。它们通过加载/存储直接访问，以避免系统调用开销。该项目还研究了超级文件的新共享语义，允许非协作进程同时安全地共享它们，同时保持崩溃恢复属性。加速超文件共享的架构支持也将被设计和评估。该项目将通过减少文件访问固有的开销来实现对持久数据的快速访问，并通过强大的共享语义实现对数据的更多并发使用。大数据工作负载在所有行业中变得越来越普遍，要么表现更好，要么需要更少的资源。所产生的软件和原型工件将使其他研究人员能够进一步增强设计，并调整工作负载以利用超文件抽象。该计划的教育和推广部分将培训下一代程序员使用持久内存，并鼓励代表性不足的学生加入下一代程序员的行列。项目将在项目期间以及之后的5年内维护数据、代码、结果、出版物和模拟器工件。每个工件将在达到足够的成熟度时发布。这些奖项将被发布在：https://sites.google.com/view/hyperfile.This。该奖项反映了美国国家科学基金会的法定使命，并通过基金会的知识价值和更广泛的影响审查标准进行评估，认为值得支持。

项目成果

CBMM: Financial Advice for Kernel Memory Managers

CBMM：内核内存管理器的财务建议

• 发表时间: 2022
• 期刊: Proceedings of the USENIX Annual Technical Conference
• 作者: Mark Mansi, Bijan Tabatabai

FBMM: Using the VFS for Extensibility in Kernel Memory Management

FBMM：使用 VFS 实现内核内存管理的可扩展性

• DOI: 10.1145/3593856.3595908
• 发表时间: 2023
• 期刊: HOTOS '23: Proceedings of the 19th Workshop on Hot Topics in Operating Systems
• 作者: Tabatabai, Bijan;Mansi, Mark;Swift, Michael M.
• 通讯作者: Swift, Michael M.

MadFS: Per-File Virtualization for Userspace Persistent Memory Filesystems

• 发表时间: 2023
• 作者: Shawn Zhong;Chenhao Ye;Guanzhou Hu;Suyan Qu;Andrea C. Arpaci-Dusseau;Remzi H. Arpaci-Dusseau;M. Swift

ASAP: A Speculative Approach to Persistence

• DOI: 10.1109/hpca53966.2022.00070
• 发表时间: 2022-04
• 期刊: 2022 IEEE International Symposium on High-Performance Computer Architecture (HPCA)
• 作者: Sujay Yadalam;Nisarg Shah;Xiangyao Yu;M. Swift

DaxVM: Stressing the Limits of Memory as a File Interface

• DOI: 10.1109/micro56248.2022.00037
• 发表时间: 2022-10
• 期刊: 2022 55th IEEE/ACM International Symposium on Microarchitecture (MICRO)
• 作者: Chloe Alverti;Vasileios Karakostas;Nikhita Kunati;G. Goumas;M. Swift