Hostping: Diagnosing Intra-host Network Bottlenecks in RDMA Servers
Kefei Liu, Zhuo Jiang, Jiao Zhang, Haoran Wei, Xiaolong Zhong, Lizhuang Tan, Tian Pan, Tao Huang
摘要
Intra-host networking was considered robust in the RDMA (Remote Direct Memory Access) network and received little attention. However, as the RNIC (RDMA NIC) line rate increases rapidly to multi-hundred gigabits, the intra-host network becomes a potential performance bottleneck for network applications. Intra-host network bottlenecks may result in degraded intra-host bandwidth and increased intra-host latency, which can severely impact network performance. However, when intra-host bottlenecks occur, they can hardly be noticed due to the lack of a monitoring system. Furthermore, existing bottleneck diagnosis mechanisms fail to diagnose intra-host bottlenecks efficiently. In this paper, we analyze the symptom of intra-host bottlenecks based on our longterm troubleshooting experience and propose Hostping, the first bottleneck monitoring and diagnosis system dedicated to intra-host networks. The core idea of Hostping is conducting loopback tests between RNICs and endpoints within the host to measure intra-host latency and bandwidth. Hostping not only discovers intra-host bottlenecks we already knew but also reveals six bottlenecks we did not notice before.
问问这篇 Paper
智能体会读完全文。
Lune 把这篇 Paper 索引到了最后一个公式,引用它的顶会 Paper 也一样。你提问,回答直接引用原文。
引用它的顶会 Paper21
- MegaScale: Scaling Large Language Model Training to More Than 10, 000 GPUsZiheng Jiang, Haibin Lin, Yinmin Zhong, Qi Huang 等NSDI 2024 · 被引用 415 次
- Characterization of Large Language Model Development in the DatacenterQinghao Hu, Zhisheng Ye, Zerui Wang, Guoteng Wang 等NSDI 2024 · 被引用 192 次
- Crux: GPU-Efficient Communication Scheduling for Deep Learning TrainingJiamin Cao, Yu Guan, Kun Qian, Jiaqi Gao 等SIGCOMM 2024 · 被引用 60 次
- Minder: Faulty Machine Detection for Large-scale Distributed Model TrainingYangtao Deng, Xiang Shi, Zhuo Jiang, Xingjian Zhang 等NSDI 2025 · 被引用 36 次
- White-Boxing RDMA with Packet-Granular Software ControlChenxingyu Zhao, Jaehong Min, Ming Liu, Arvind KrishnamurthyNSDI 2025 · 被引用 28 次
它引用的顶会 Paper7
- A Unified Architecture for Accelerating Distributed DNN Training in Heterogeneous GPU/CPU ClustersYimin Jiang, Yibo Zhu, Chang Lan, Bairen Yi 等OSDI 2020 · 被引用 390 次
- When Cloud Storage Meets RDMAYixiao Gao, Qiang Li, Lingbo Tang, Yongqing Xi 等NSDI 2021 · 被引用 228 次
- Reexamining Direct Cache Access to Optimize I/O Intensive Applications for Multi-hundred-gigabit NetworksAlireza Farshin, Amir Roozbeh, Gerald Q. Maguire Jr., Dejan KosticUSENIX ATC 2020 · 被引用 88 次
- Collie: Finding Performance Anomalies in RDMA SubsystemsXinhao Kong, Yibo Zhu, Huaping Zhou, Zhuo Jiang 等NSDI 2022 · 被引用 86 次
- How to diagnose nanosecond network latencies in rich end-host stacksRoni Haecki, Radhika Niranjan Mysore, Lalith Suresh, Gerd Zellweger 等NSDI 2022 · 被引用 58 次
相关 Paper
- R-Pingmesh: A Service-Aware RoCE Network Monitoring and Diagnostic SystemKefei Liu, Zhuo Jiang, Jiao Zhang, Shixian Guo 等SIGCOMM 2024 · 被引用 21 次
- Understanding the Host NetworkMidhul Vuppalapati, Saksham Agarwal, Henry Schuh, Baris Kasikci 等SIGCOMM 2024 · 被引用 25 次
- Towards Lightweight Traffic Forecasting in RDMA Networks: Design and ApplicationCheng Yang, Xiaoning Zhang, Bodong Yan, Sun Xu 等INFOCOM 2025 · 被引用 1 次
- INSERT: In-Network Stateful End-to-End RDMA TelemetryHyunseok Chang, Walid A. Hanafy, Sarit Mukherjee, Limin WangINFOCOM 2024 · 被引用 3 次
- Software-based Live Migration for RDMAXiaoyu Li, Ran Shu, Yongqiang Xiong, Fengyuan RenSIGCOMM 2025 · 被引用 4 次
