[Resource Topic] 2025/167: Wiretapping LLMs: Network Side-Channel Attacks on Interactive LLM Services

Welcome to the resource topic for 2025/167

Title:
Wiretapping LLMs: Network Side-Channel Attacks on Interactive LLM Services

Authors: Mahdi Soleimani, Grace Jia, In Gim, Seung-seob Lee, Anurag Khandelwal

Abstract:

Recent server-side optimizations like speculative decoding significantly enhance the interactivity and resource efficiency of Large Language Model (LLM) services. However, we show that these optimizations inadvertently introduce new side-channel vulnerabilities through network packet timing and size variations that tend to be input-dependent. Network adversaries can leverage these side channels to learn sensitive information contained in \emph{encrypted} user prompts to and responses from public LLM services.

This paper formalizes the security implications using a novel indistinguishability framework and introduces a novel attack that establishes the insecurity of real-world LLM services with streaming APIs under our security framework.

Our proposed attack effectively deconstructs encrypted network packet traces to reveal the sizes of underlying LLM-generated tokens and whether the tokens were generated with or without certain server-side optimizations. Our attack can accurately predict private attributes in real-world privacy-sensitive LLM applications in medicine and finance with 7192\% accuracy on an open-source vLLM service and 5090\% accuracy on the commercial ChatGPT service. Finally, we show that solutions that hide these side channels to different degrees expose a tradeoff between security and performance — specifically, interactivity and network bandwidth overheads.

ePrint: https://eprint.iacr.org/2025/167

See all topics related to this paper.

Feel free to post resources that are related to this paper below.

Example resources include: implementations, explanation materials, talks, slides, links to previous discussions on other websites.

For more information, see the rules for Resource Topics .