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

coucouArray ( [post_type] => ndss-paper [post_status] => publish [posts_per_page] => 4 [orderby] => rand [tax_query] => Array ( [0] => Array ( [taxonomy] => category [field] => id [terms] => Array ( [0] => 118 ) ) ) [post__not_in] => Array ( [0] => 20083 ) )

DiStefano: Decentralized Infrastructure for Sharing Trusted Encrypted Facts and...

Sofia Celi (Brave Software), Alex Davidson (NOVA LINCS & Universidade NOVA de Lisboa), Hamed Haddadi (Imperial College London & Brave Software), Gonçalo Pestana (Hashmatter), Joe Rowell (Information Security Group, Royal Holloway, University of London)

Read More

Privacy-Preserving Data Deduplication for Enhancing Federated Learning of Language...

Aydin Abadi (Newcastle University), Vishnu Asutosh Dasu (Pennsylvania State University), Sumanta Sarkar (University of Warwick)

Read More

SHAFT: Secure, Handy, Accurate and Fast Transformer Inference

Andes Y. L. Kei (Chinese University of Hong Kong), Sherman S. M. Chow (Chinese University of Hong Kong)

Read More

Privacy Starts with UI: Privacy Patterns and Designer Perspectives in UI/UX Practice

Anxhela Maloku (Technical University of Munich), Alexandra Klymenko (Technical University of Munich), Stephen Meisenbacher (Technical University of Munich), Florian Matthes (Technical University of Munich)

Vision: Profiling Human Attackers: Personality and Behavioral Patterns in Deceptive Multi-Stage CTF Challenges

Khalid Alasiri (School of Computing and Augmented Intelligence Arizona State University), Rakibul Hasan (School of Computing and Augmented Intelligence Arizona State University)

From Underground to Mainstream Marketplaces: Measuring AI-Enabled NSFW Deepfakes on Fiverr

Mohamed Moustafa Dawoud (University of California, Santa Cruz), Alejandro Cuevas (Princeton University), Ram Sundara Raman (University of California, Santa Cruz)