Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
121 tokens/sec
GPT-4o
9 tokens/sec
Gemini 2.5 Pro Pro
47 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Beyond the Bridge: Contention-Based Covert and Side Channel Attacks on Multi-GPU Interconnect (2404.03877v2)

Published 5 Apr 2024 in cs.CR

Abstract: High-speed interconnects, such as NVLink, are integral to modern multi-GPU systems, acting as a vital link between CPUs and GPUs. This study highlights the vulnerability of multi-GPU systems to covert and side channel attacks due to congestion on interconnects. An adversary can infer private information about a victim's activities by monitoring NVLink congestion without needing special permissions. Leveraging this insight, we develop a covert channel attack across two GPUs with a bandwidth of 45.5 kbps and a low error rate, and introduce a side channel attack enabling attackers to fingerprint applications through the shared NVLink interconnect.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (14)
  1. Y. Zhang, D. Pandey, D. Wu, T. Kundu, R. Li, and T. Shu, “Accuracy-Constrained Efficiency Optimization and GPU Profiling of CNN Inference for Detecting Drainage Crossing Locations,” Workshops of ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis (SC-W), 2023.
  2. Y. Zhang, C. Slocum, J. Chen, and N. Abu-Ghazaleh, “It’s all in your head(set): Side-channel attacks on AR/VR systems,” in USENIX Security, 2023.
  3. Y. Li, Z. Li, W. Yang, and C. Liu, “Rt-lm: Uncertainty-aware resource management for real-time inference of language models,” arXiv preprint arXiv:2309.06619, 2023.
  4. Z. Lai, J. Wu, S. Chen, Y. Zhou, A. Hovakimyan, and N. Hovakimyan, “Language models are free boosters for biomedical imaging tasks,” arXiv preprint arXiv:2403.17343, 2024.
  5. M. Tan, J. Wan, Z. Zhou, and Z. Li, “Invisible probe: Timing attacks with pcie congestion side-channel,” in 2021 IEEE Symposium on Security and Privacy (SP).   IEEE, 2021, pp. 322–338.
  6. S. B. Dutta, H. Naghibijouybari, N. Abu-Ghazaleh, A. Marquez, and K. Barker, “Leaky buddies: Cross-component covert channels on integrated cpu-gpu systems,” in 2021 ACM/IEEE 48th Annual International Symposium on Computer Architecture (ISCA).   IEEE, 2021, pp. 972–984.
  7. M. Side, F. Yao, and Z. Zhang, “Lockeddown: Exploiting contention on host-gpu pcie bus for fun and profit,” in 2022 IEEE 7th European Symposium on Security and Privacy (EuroS&P).   IEEE, 2022, pp. 270–285.
  8. Nvidia, “NVLink and NVSwitch,” https://www.nvidia.com/en-us/data-center/nvlink/, 2014.
  9. J. Choquette, O. Giroux, and D. Foley, “Volta: Performance and programmability,” Ieee Micro, vol. 38, no. 2, pp. 42–52, 2018.
  10. H. Naghibijouybari, A. Neupane, Z. Qian, and N. Abu-Ghazaleh, “Rendered insecure: Gpu side channel attacks are practical,” in Proceedings of the 2018 ACM SIGSAC conference on computer and communications security, 2018, pp. 2139–2153.
  11. J. Wei, Y. Zhang, Z. Zhou, Z. Li, and M. A. Al Faruque, “Leaky dnn: Stealing deep-learning model secret with gpu context-switching side-channel,” in 2020 50th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN).   IEEE, 2020, pp. 125–137.
  12. P. Eastman, J. Swails, J. D. Chodera, R. T. McGibbon, Y. Zhao, K. A. Beauchamp, L.-P. Wang, A. C. Simmonett, M. P. Harrigan, C. D. Stern et al., “Openmm 7: Rapid development of high performance algorithms for molecular dynamics,” PLoS computational biology, vol. 13, no. 7, p. e1005659, 2017.
  13. S. B. Dutta, H. Naghibijouybari, A. Gupta, N. Abu-Ghazaleh, A. Marquez, and K. Barker, “Spy in the gpu-box: Covert and side channel attacks on multi-gpu systems,” in Proceedings of the 50th Annual International Symposium on Computer Architecture, 2023, pp. 1–13.
  14. Y. Zhang, R. Yasaei, H. Chen, Z. Li, and M. A. Al Faruque, “Stealing neural network structure through remote fpga side-channel analysis,” IEEE Transactions on Information Forensics and Security, vol. 16, pp. 4377–4388, 2021.
Citations (4)

Summary

We haven't generated a summary for this paper yet.

X Twitter Logo Streamline Icon: https://streamlinehq.com