SCALE: Tackling Communication Bottlenecks in Confidential Distributed Machine Learning

计算机科学 保密 人工智能 机器学习 钥匙(锁) 匹配(统计) 分布式学习 计算机安全 特征(语言学) 瓶颈
作者
Joongun Park,Yongqin Wang,Huan Xu,Hanjiang Wu,Mengyuan Li,Tushar Krishna
出处
期刊:Proceedings [Institute of Electrical and Electronics Engineers]
卷期号:: 1-14
标识
DOI:10.1109/hpca68181.2026.11408582
摘要

Machine Learning (ML) has become a cornerstone in numerous applications, creating the need for secure and efficient distributed ML frameworks. However, maintaining data privacy in these systems poses significant challenges, particularly in distributed environments where user data and model parameters must frequently be transmitted between GPUs. Confidential GPU computing technologies, such as NVIDIA's Confidential Computing (CC) mode, offer hardware-based enterprise solutions designed to protect ML workloads in untrusted environments (e.g., public clouds). These technologies leverage heterogeneous systems that combine Confidential Virtual Machines (CVMs) with GPU-based Trusted Execution Environments (TEEs). Nevertheless, confidential computing introduces considerable performance overhead due to its complex heterogeneous architecture and the high-throughput data flows required across TEE security boundaries. For example, encrypted communication occurs both between CVMs and GPU TEEs, and among multiple GPU TEEs, resulting in significant latency compared to native PCIe or high-speed interconnects such as NVLink. Our extensive evaluation shows that these overheads become particularly severe during collective communication operations, which suffer from encryption-induced delays that negatively impact end-to-end training performance. To address this, we propose a co-encryption design that leverages underutilized GPU resources, optimizes encryption and authentication, and introduces a communication algorithm tailored for confidential settings. We evaluate our design using real ML workloads and execution traces collected from four HGX H100/H200 clusters. While CC mode was not available on current NVIDIA software stacks, we incorporate encryptionaware modeling based on hardware specifications to estimate secure communication overheads. Our results demonstrate a$\mathbf{4 0 - 7 0 \%}$reduction in communication-related security costs.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
慕青的应助被鱼鱼鱼采纳,获得10
刚刚
顾矜的应助被suoyu采纳,获得10
2秒前
Firsterchao发布了新的文献求助10
2秒前
sjyu1985完成签到 ,获得积分10
2秒前
酷酷的悒发布了新的文献求助10
3秒前
LYU完成签到,获得积分10
4秒前
milz发布了新的文献求助10
4秒前
传奇3的应助被畅快访蕊采纳,获得10
4秒前
4秒前
西柏完成签到,获得积分10
4秒前
5秒前
科研通AI6.3的应助被Zoey采纳,获得10
5秒前
科研通AI6.2的应助被暮光之城采纳,获得10
5秒前
7秒前
刘芋叶发布了新的文献求助10
8秒前
打打的应助被张朝阳采纳,获得10
9秒前
华仔的应助被Alive采纳,获得10
10秒前
刘刘发布了新的文献求助10
11秒前
健壮的南琴完成签到,获得积分20
11秒前
程君完成签到,获得积分10
11秒前
YAN完成签到 ,获得积分10
12秒前
Nole的应助被观山青采纳,获得10
12秒前
文艺雁梅发布了新的文献求助10
12秒前
13秒前
14秒前
赏金猎人John_Wang完成签到,获得积分10
14秒前
14秒前
lmy完成签到,获得积分10
15秒前
还等雨落的应助被清风徐来采纳,获得10
15秒前
15秒前
15秒前
15秒前
123发布了新的文献求助10
15秒前
15秒前
无极微光的应助被难过的谷芹采纳,获得20
16秒前
浅笑成风完成签到,获得积分10
17秒前
GIN的应助被积木采纳,获得10
17秒前
LynShen发布了新的文献求助10
17秒前
18秒前
畅快访蕊发布了新的文献求助10
19秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
拉伸高分子熔体中缠结行为的模拟和小角中子散射研究 2000
Agyptische Geschichte der 21.30. Dynastie 510
高分子基功能界面水蒸发材料与应用研究 500
外场增强的水汽发电研究 500
基于结构调控的功能化水诱导发电材料与器件研究 500
基于界面调控的高效湿气发电材料与器件研究 500
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 工程类 有机化学 化学工程 生物化学 物理 复合材料 内科学 催化作用 光电子学 人工智能 细胞生物学 无机化学 物理化学 数学
热门帖子
关注 科研通微信公众号,转发送积分 8192723
求助须知:如何正确求助?哪些是违规求助? 9579144
关于积分的说明 21863903
捐赠科研通 7792268
什么是DOI,文献DOI怎么找? 3437013
关于科研通互助平台的介绍 2560633
邀请新用户注册赠送积分活动 2483864