paperbot · PL 论文追踪

全部论文 待读 收藏

P2 sIOPMP: Scalable and Efficient I/O Protection for TEEs

sIOPMP:面向 TEE 的可扩展高效 I/O 保护
ASPLOS Proceedings of the 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 22024匹配度 1/10 · 该论文研究 TEE 的 I/O 物理内存保护机制,属于计算机体系结构与硬件安全方向,与用户关注的 PL/AI 交叉、静态分析、Agentic 程序、浮点程序分析、贝叶斯程序分析等方向基本无关。引用 17 inbox
Erhu Feng, Dahu Feng, Dong Du, Yubin Xia, Wenbin Zheng, Siqi Zhao, Haibo Chen
crossref: 2026-09-12dblp: 2026-09-12

AI 速览

一句话:提出 sIOPMP,一种面向 TEE 的可扩展 I/O 物理内存保护机制,以低开销实现大量设备隔离并提升吞吐。

TEE(如 Intel SGX/TDX、AMD SEV-SNP、ARM TrustZone/CCA)已被广泛采用,但通常未将 I/O 隔离(例如防御恶意 DMA 请求)作为一等需求,导致 I/O 密集负载性能下降。传统 IOMMU 或软件 I/O 方案可使吞吐至少降低 20%,原因是 I/O 设备的隔离需求不同于 CPU。本文提出 sIOPMP(可扩展 I/O 物理内存保护),包含三项关键设计:基于多级树的检查器,支持超过 1000 个硬件区域;将设备分为热设备和冷设备,并通过可挂载条目支持无限设备;提出重映射机制,在动态 I/O 负载下切换设备的热/冷状态。评估表明,sIOPMP 在基准测试和真实负载中开销可忽略,相比基于 IOMMU 的机制或 TEE 中采用的软件 I/O,网络吞吐提升 20%~38%。

问题
TEE 通常未将 I/O 隔离作为一等需求,难以防御恶意 DMA 请求;传统 IOMMU 或软件 I/O 方案会使 I/O 密集负载吞吐至少下降 20%,因为 I/O 设备的隔离需求与 CPU 不同。
方法
提出 sIOPMP,采用多级树检查器支持超过 1000 个硬件区域,将设备分为热/冷两类并以可挂载条目支持无限设备,同时通过重映射机制在动态负载下切换设备状态。
评估
在基准测试和真实负载上评估,显示性能开销可忽略,并且相比基于 IOMMU 的机制或软件 I/O,网络吞吐提升 20%~38%。
原文摘要(Abstract)

Trusted Execution Environments (TEEs), like Intel SGX/TDX, AMD SEV-SNP, ARM TrustZone/CCA, have been widely adopted in prevailing architectures. However, these TEEs typically do not consider I/O isolation (e.g., defending against malicious DMA requests) as a first-class citizen, which may degrade the I/O performance. Traditional methods like using IOMMU or software I/O can degrade throughput by at least 20% for I/O intensive workloads. The main reason is that the isolation requirements for I/O devices differ from CPU ones. This paper proposes a novel I/O isolation mechanism for TEEs, named sIOPMP (scalable I/O Physical Memory Protection), with three key features. First, we design a Multi-stage-Tree-based checker, supporting more than 1,000 hardware regions. Second, we classify the devices into hot and cold, and support unlimited devices with the mountable entry. Third, we propose a remapping mechanism to switch devices between hot and cold status for dynamic I/O workloads. Evaluation results show that sIOPMP introduces only negligible performance overhead for both benchmarks and real-world workloads, and improves 20% ~ 38% network throughput compared with IOMMU-based mechanisms or software I/O adopted in TEEs.

链接与引用

DOI 原文 · DBLP

BibTeX
@inproceedings{FengF0XZZ024,
  title = {sIOPMP: Scalable and Efficient I/O Protection for TEEs},
  author = {Erhu Feng and Dahu Feng and Dong Du and Yubin Xia and Wenbin Zheng and Siqi Zhao and Haibo Chen},
  booktitle = {ASPLOS},
  year = {2024},
  doi = {10.1145/3620665.3640378}
}