Yuri Gbur (Technische Universität Berlin), Florian Tschorsch (Technische Universität Berlin)

The QUIC protocol is gaining more and more traction through its recent standardization and the rising interest by various big tech companies, developing new implementations. QUIC promises to make security and privacy a first-class citizen; yet, challenging these claims is of utmost importance. To this end, this paper provides an initial analysis of client-side request forgery attacks that directly emerge from the QUIC protocol design and not from common vulnerabilities. In particular, we investigate three request forgery attack modalities with respect to their capabilities to be used for protocol impersonation and traffic amplification. We analyze the controllable attack space of the respective protocol messages and demonstrate that one of the attack modalities can indeed be utilized to impersonate other UDP-based protocols, e.g., DNS requests. Furthermore, we identify traffic amplification vectors. Although the QUIC protocol specification states anti-amplification limits, our evaluation of 13 QUIC server implementations shows that in some cases these mitigations are missing or insufficiently implemented. Lastly, we propose mitigation approaches for protocol impersonation and discuss ambiguities in the specification.

View More Papers

Him of Many Faces: Characterizing Billion-scale Adversarial and Benign...

Shujiang Wu (Johns Hopkins University), Pengfei Sun (F5, Inc.), Yao Zhao (F5, Inc.), Yinzhi Cao (Johns Hopkins University)

Read More

Let Me Unwind That For You: Exceptions to Backward-Edge...

Victor Duta (Vrije Universiteit Amsterdam), Fabian Freyer (University of California San Diego), Fabio Pagani (University of California, Santa Barbara), Marius Muench (Vrije Universiteit Amsterdam), Cristiano Giuffrida (Vrije Universiteit Amsterdam)

Read More

QPEP in the Real World: A Testbed for Secure...

Julian Huwyler (ETH Zurich), James Pavur (University of Oxford), Giorgio Tresoldi and Martin Strohmeier (Cyber-Defence Campus) Presenter: Martin Strohmeier

Read More

BARS: Local Robustness Certification for Deep Learning based Traffic...

Kai Wang (Tsinghua University), Zhiliang Wang (Tsinghua University), Dongqi Han (Tsinghua University), Wenqi Chen (Tsinghua University), Jiahai Yang (Tsinghua University), Xingang Shi (Tsinghua University), Xia Yin (Tsinghua University)

Read More