CNS Core: Medium: Secure Hardware Virtualization Contexts for Ephemeral Cloud Computing

CNS 核心：中：用于临时云计算的安全硬件虚拟化上下文

• 批准号: 2107307
• 负责人: Todd Mowry
• 金额: $ 60万
• 依托单位: Carnegie-Mellon University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-06-01 至 2024-05-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2107307&HistoricalAwards=false
• 关键词: CNS; Core; Medium; Secure; Hardware

Lightweight virtualization technologies such as containers have enabled emerging computing paradigms of microservices and serverless computing with thousands of concurrent users sharing each server. Unfortunately, current compiler, hardware, and operating systems were optimized for the steady state of long running applications instead of the very short-lived execution of functions. Especially in the current era of multi-core processors, existing context management mechanisms impose significant performance overheads. Furthermore, serverless environments are dominated by ephemeral state (e.g., intermediate computations) and containers that are discarded after function invocation is complete. Nevertheless, existing system stacks do not take into account the ephemeral nature of serverless computing. The goal of this project is to design new hardware and operating system execution contexts in synergy with compiler optimizations to match the needs of function computation. This work will build new compiler techniques for function analysis to identify the data needs of ephemeral execution. In addition, this project will explore a new cross-stack interface that allows the communication and ephemeral state tracking at the operating system and hardware layers. Finally, this work will explore the design of hardware support for managing compiler-identified ephemeral data through novel hardware contexts purposefully built for function execution that will aim to minimize data movement and energy consumption.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.

容器等轻量级虚拟化技术支持微服务和无服务器计算的新兴计算范式，其中数千个并发用户共享每个服务器。不幸的是，当前的编译器、硬件和操作系统都是针对长时间运行的应用程序的稳定状态进行优化的，而不是针对非常短暂的函数执行进行优化。特别是在当前的多核处理器时代，现有的上下文管理机制带来了巨大的性能开销。此外，无服务器环境由短暂状态（例如，中间计算）和在函数调用完成后丢弃的容器主导。然而，现有的系统堆栈并没有考虑到无服务器计算的短暂性。这个项目的目标是设计新的硬件和操作系统执行上下文，与编译器优化协同，以匹配函数计算的需要。这项工作将为函数分析构建新的编译器技术，以确定临时执行的数据需求。此外，该项目将探索一个新的跨堆栈接口，该接口允许在操作系统和硬件层进行通信和短暂状态跟踪。最后，本工作将探索硬件支持的设计，通过专门为函数执行构建的新颖硬件上下文来管理编译器识别的临时数据，旨在最大限度地减少数据移动和能耗。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

TMO: transparent memory offloading in datacenters

• DOI: 10.1145/3503222.3507731
• 发表时间: 2022-02
• 期刊: Proceedings of the 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems
• 作者: Johannes Weiner;Niket Agarwal;Dan Schatzberg;Leon Yang;Hao Wang-;Blaise Sanouillet;Bikash Sharma;Tejun Heo;M. Jain;Chunqiang Tang;Dimitrios Skarlatos

Memory-Efficient Hashed Page Tables

• DOI: 10.1109/hpca56546.2023.10071061
• 发表时间: 2023-02
• 期刊: 2023 IEEE International Symposium on High-Performance Computer Architecture (HPCA)
• 作者: Jovan Stojkovic;Namrata Mantri;Dimitrios Skarlatos;Tianyi Xu;J. Torrellas

IOCost: block IO control for containers in datacenters

IOCost：数据中心容器的块IO控制

• DOI: 10.1145/3503222.3507727
• 发表时间: 2022
• 期刊: Proceedings of the 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems
• 作者: Heo, Tejun;Schatzberg, Dan;Newell, Andrew;Liu, Song;Dhakshinamurthy, Saravanan;Narayanan, Iyswarya;Bacik, Josef;Mason, Chris;Tang, Chunqiang;Skarlatos, Dimitrios
• 通讯作者: Skarlatos, Dimitrios

Contiguitas: The Pursuit of Physical Memory Contiguity in Datacenters

• DOI: 10.1145/3579371.3589079
• 发表时间: 2023-06
• 期刊: Proceedings of the 50th Annual International Symposium on Computer Architecture
• 作者: Kaiyang Zhao;Kaiwen Xue;Ziqi Wang-;Dan Schatzberg;Leon Yang;Antonis Manousis;Johannes Weiner;Rik Van Riel;Bikash Sharma;Chunqiang Tang;Dimitrios Skarlatos

Parallel virtualized memory translation with nested elastic cuckoo page tables

使用嵌套弹性布谷鸟页表进行并行虚拟化内存转换

• DOI: 10.1145/3503222.3507720
• 发表时间: 2022
• 期刊: Proceedings of the 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems
• 作者: Stojkovic, Jovan;Skarlatos, Dimitrios;Kokolis, Apostolos;Xu, Tianyin;Torrellas, Josep
• 通讯作者: Torrellas, Josep