Collaborative Research: CNS Core: Medium: Understanding and Strengthening Memory Security for Non-Volatile Memory

合作研究：CNS 核心：中：理解和加强非易失性内存的内存安全性

• 批准号: 2106893
• 负责人: Felix Xiaozhu Lin
• 金额: $ 40万
• 依托单位: University of Virginia Main Campus
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-10-01 至 2024-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2106893&HistoricalAwards=false
• 关键词: Collaborative; Research; CNS; Core; Medium

Memory safety is essential. Despite decades of research, unauthorized memory reads and writes are still among the most common security attacks. The emerging persistent memory (PM) amplifies the importance of strong memory protections. As a promising supplement or substitute of DRAM as main memory, PM offers higher density, better scaling potential, lower idle power, and non-volatility, while retaining byte addressability and random accessibility. But PM poses some principled differences for protection in comparison to traditional memory in several ways: PM keeps persistent data so data corruption is persistent and costlier; there is no obvious way to recover from data corruption; due to PM data being long lived, its content and structure and hence the impact of data corruption are likely shared by one or more applications over time. This project aims to significantly advance the understanding of PM memory protection, establish the foundation for persistence-conscious memory protection, create a set of novel techniques for protecting PM memory, and for detection, recovery, and diagnostic of PM corruption. The growth of computer performance that has enabled many scientific discoveries and inventions, is slowing down due to the looming end of Moore’s Law. Persistent memory (PM) is one of promising post-Moore’s Law technology that is expected to get into the mainstream computing devices, from internet of things devices all the way to the cloud. This research anticipates security vulnerabilities of the use of PM and attempts to mitigate potential security breaches that could cause tremendous loss to the industry, defense, scientific research, health, and many other domains that will be relying on future computing devices.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.

内存安全至关重要。尽管经过几十年的研究，未经授权的内存读写仍然是最常见的安全攻击之一。新出现的永久记忆(PM)放大了强大记忆保护的重要性。作为DRAM作为主存储器的一种有前景的补充或替代，PM具有更高的密度、更好的扩展潜力、更低的空闲功耗和非易失性，同时保持了字节可寻址和随机访问。但与传统内存相比，PM在保护方面存在一些原则性差异：PM保留持久数据，因此数据损坏是持久的，成本更高；没有明显的方法可以从数据损坏中恢复；由于PM数据的生命期很长，其内容和结构以及数据损坏的影响可能会随着时间的推移由一个或多个应用程序共享。该项目旨在显著提高对PM内存保护的理解，为持久性记忆保护奠定基础，创造一套新的PM内存保护技术，并用于PM腐败的检测、恢复和诊断。计算机性能的增长使许多科学发现和发明成为可能，但由于摩尔定律即将结束，增长速度正在放缓。永久存储器(PM)是一种很有前途的后摩尔定律技术，有望进入从物联网设备到云计算的主流计算设备。这项研究预测了PM使用的安全漏洞，并试图缓解潜在的安全漏洞，这些漏洞可能会给依赖于未来计算设备的工业、国防、科学研究、健康和许多其他领域造成巨大损失。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势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

Efficient NLP Model Finetuning via Multistage Data Filtering

• DOI: 10.24963/ijcai.2023/455
• 发表时间: 2022-07
• 作者: Ouyang Xu;S. Ansari;F. Lin;Yangfeng Ji

STI: Turbocharge NLP Inference at the Edge via Elastic Pipelining

• DOI: 10.1145/3575693.3575698
• 发表时间: 2022-07
• 期刊: Proceedings of the 28th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2
• 作者: Liwei Guo;Wonkyo Choe;F. Lin

Safe and Practical GPU Computation in TrustZone

• DOI: 10.1145/3552326.3567483
• 发表时间: 2023-05
• 期刊: Proceedings of the Eighteenth European Conference on Computer Systems
• 作者: Heejin Park;F. Lin

Towards Out-of-core Neural Networks on Microcontrollers

• DOI: 10.1109/sec54971.2022.00008
• 发表时间: 2022-12
• 期刊: 2022 IEEE/ACM 7th Symposium on Edge Computing (SEC)
• 作者: Hongyu Miao;F. Lin

Incremental perception on real time 3D data

• DOI: 10.1145/3508396.3512875
• 发表时间: 2022-03
• 期刊: Proceedings of the 23rd Annual International Workshop on Mobile Computing Systems and Applications
• 作者: A. Sarker;F. Lin