Haohuang Wen (The Ohio State University), Phillip Porras (SRI International), Vinod Yegneswaran (SRI International), Ashish Gehani (SRI International), Zhiqiang Lin (The Ohio State University)

Over the past several years, the mobile security community has discovered a wide variety of exploits against link and session-establishment protocols. These exploits can be implemented on software-defined radios (SDRs) that disrupt, spoof, or flood layer-3 (L3) messages to compromise security and privacy, which still apply to the latest 5G mobile network standard. Interestingly, unlike the prior generations of closed (proprietary) mobile network infrastructures, 5G networks are migrating toward a more intelligent and open-standards-based fully interoperable mobile architecture, called Open RAN or O-RAN. The implications of transitioning mobile infrastructures to a software-defined architectural abstraction are quite significant to the INFOSEC community, as it allows us to extend the mobile data plane and control plane with security-focused protocol auditing services and exploit detection. Based on this design, we present 5G-SPECTOR, the first comprehensive framework for detecting the wide spectrum of L3 protocol exploits on O-RAN. It features a novel security audit stream called MOBIFLOW that transfers fine-grained cellular network telemetry, and a programmable control-plane xApp called MOBIEXPERT. We present an extensible prototype of 5G-SPECTOR which can detect 7 types of cellular attacks in real time. We also demonstrate its scalability to 11 unknown attacks as well as 31 real-world cellular traces, with effective performance (high accuracy, no false alarms) and low (<2% CPU, <100 MB memory) overhead.

View More Papers

TALISMAN: Tamper Analysis for Reference Monitors

Frank Capobianco (The Pennsylvania State University), Quan Zhou (The Pennsylvania State University), Aditya Basu (The Pennsylvania State University), Trent Jaeger (The Pennsylvania State University, University of California, Riverside), Danfeng Zhang (The Pennsylvania State University, Duke University)

Read More

A Comparative Analysis of Difficulty Between Log and Graph-Based...

Matt Jansen, Rakesh Bobba, Dave Nevin (Oregon State University)

Read More

WIP: Towards a Certifiably Robust Defense for Multi-label Classifiers...

Dennis Jacob, Chong Xiang, Prateek Mittal (Princeton University)

Read More

CrowdGuard: Federated Backdoor Detection in Federated Learning

Phillip Rieger (Technical University of Darmstadt), Torsten Krauß (University of Würzburg), Markus Miettinen (Technical University of Darmstadt), Alexandra Dmitrienko (University of Würzburg), Ahmad-Reza Sadeghi (Technical University of Darmstadt)

Read More