Guanlong Wu (Southern University of Science and Technology), Zheng Zhang (ByteDance Inc.), Yao Zhang (ByteDance Inc.), Weili Wang (Southern University of Science and Technolog), Jianyu Niu (Southern University of Science and Technolog), Ye Wu (ByteDance Inc.), Yinqian Zhang (Southern University of Science and Technology (SUSTech))

Large Language Models (LLMs), which laid the groundwork for Artificial General Intelligence (AGI), have recently gained significant traction in academia and industry due to their disruptive applications. In order to enable scalable applications and efficient resource management, various multi-tenant LLM serving frameworks have been proposed, in which the LLM caters to the needs of multiple users simultaneously. One notable mechanism in recent works, such as SGLang and vLLM, is sharing the Key-Value (KV) cache for identical token sequences among multiple users, saving both memory and computation. This paper presents the first investigation on security risks
associated with multi-tenant LLM serving. We show that the state-of-the-art mechanisms of KV cache sharing may lead to new side channel attack vectors, allowing unauthorized reconstruction
of user prompts and compromising sensitive user information among mutually distrustful users. Specifically, we introduce our attack, PROMPTPEEK, and apply it to three scenarios where the
adversary, with varying degrees of prior knowledge, is capable of reverse-engineering prompts from other users. This study underscores the need for careful resource management in multi-tenant LLM serving and provides critical insights for future security enhancement.

View More Papers

Logical Maneuvers: Detecting and Mitigating Adversarial Hardware Faults in...

Fatemeh Khojasteh Dana, Saleh Khalaj Monfared, Shahin Tajik (Worcester Polytechnic Institute)

Read More

Enhancing Security in Third-Party Library Reuse – Comprehensive Detection...

Shangzhi Xu (The University of New South Wales), Jialiang Dong (The University of New South Wales), Weiting Cai (Delft University of Technology), Juanru Li (Feiyu Tech), Arash Shaghaghi (The University of New South Wales), Nan Sun (The University of New South Wales), Siqi Ma (The University of New South Wales)

Read More

Black-box Membership Inference Attacks against Fine-tuned Diffusion Models

Yan Pang (University of Virginia), Tianhao Wang (University of Virginia)

Read More

Vision: Comparison of AI-assisted Policy Development Between Professionals and...

Rishika Thorat (Purdue University), Tatiana Ringenberg (Purdue University)

Read More