StormShield: Fingerprint-Based Detection and Mitigation of RRC Signaling Storms in O-RAN 5G RANs CCS: Networks Denial-of-service attacks Conference: Proceedings of the 19th ACM Conference on Security and Privacy in Wireless and Mobile Networks; June 30-July 03, 2026; Saarbrücken, Germany Proceedings of the 19th ACM Conference on Security and Privacy in Wireless and Mobile Networks (WiSec ’26), June 30-July 03, 2026, Saarbrücken, Germany DOI: 10.1145/3765613.3811685 ISBN: 979-8-4007-2201-1/2026/06
Noemi Giustini, Andrea Lacava, Leonardo Bonati, Stefano Maxenti, Michele Polese, Tommaso Melodia, Francesca Cuomo · 2026
Read StormShield: Fingerprint-Based Detection and Mitigation of RRC Signaling Storms in O-RAN 5G RANs CCS: Networks Denial-of-service attacks Conference: Proceedings of the 19th ACM Conference on Security and Privacy in Wireless and Mobile Networks; June 30-July 03, 2026; Saarbrücken, Germany Proceedings of the 19th ACM Conference on Security and Privacy in Wireless and Mobile Networks (WiSec ’26), June 30-July 03, 2026, Saarbrücken, Germany DOI: 10.1145/3765613.3811685 ISBN: 979-8-4007-2201-1/2026/06 by Noemi Giustini, Andrea Lacava, Leonardo Bonati, Stefano Maxenti, Michele Polese, Tommaso Melodia, Francesca Cuomo (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-14032.txt.
Introduction
Giustini and six co-authors present StormShield, a fingerprint-based defense against RRC signaling storms in O-RAN 5G networks. In such an attack, a rogue user device floods the base station with signaling, draining its resources and blocking legit devices from connecting. Older defenses mostly just detect attacks, rest on simulations alone, and cannot tell heavy load from attack or handle device motion. The paper first tests whether the attack works on an open-source radio testbed.
Catalog record
| CREATOR | Noemi Giustini, Andrea Lacava, Leonardo Bonati, Stefano Maxenti, Michele Polese, Tommaso Melodia, Francesca Cuomo |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 55761 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.14032
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
StormShield: Fingerprint-Based Detection and Mitigation of RRC Signaling Storms in O-RAN 5G RANs CCS: Networks Denial-of-service attacks Conference: Proceedings of the 19th ACM Conference on Security and Privacy in Wireless and Mobile Networks; June 30-July 03, 2026; Saarbrücken, Germany Proceedings of the 19th ACM Conference on Security and Privacy in Wireless and Mobile Networks (WiSec ’26), June 30-July 03, 2026, Saarbrücken, Germany DOI: 10.1145/3765613.3811685 ISBN: 979-8-4007-2201-1/2026/06 Noemi Giustini, Andrea Lacava, Leonardo Bonati, Stefano Maxenti, Michele Polese, Tommaso Melodia, Francesca Cuomo Abstract . 5G networks provide low-latency, high throughput, and massive connectivity, yet the control plane remains exposed to several security threats. Among the most common and impactful threats are Denial-of-Service (DoS) attacks, with Radio Resource Control (RRC) signaling storms being particularly effective and difficult to mitigate. In this attack, a malicious User Equipment (UE) aims to exhaust Next Generation Node Base (gNB) resources, preventing legitimate UEs from establishing a connection. Existing defenses are typically limited to detection, only evaluated through numerical simulations, and cannot discern between high-load network conditions and attacks. Most of them also assume static setups and do not take mobility into account. In this paper, we first evaluate the feasibility of the signaling storm attack by using the OAI 5G protocol stack. Then, we propose StormShield, a signaling storm attack detection and mitigation technique implemented as an xApp on an O-RAN Near-Real-Time (near-RT) RAN Intelligent Controller (RIC) . It fingerprints and blocks Malicious UEs (MUEs) before gNB resources are exhausted. We prototyped our solution on an Over-The-Air (OTA) testbed with OAI , NVIDIA Aerial, and two different gNB setups. The first one leverages an USRP X410 Software-defined Radio (SDR) with 8.1 functional split; the second a commercial Foxconn Radio Unit (RU) with 7.2 functional split. Our experimental evaluation demonstrates that StormShield effectively prevents gNB resource exhaustion, identifying and blocking MUEs with an average detection accuracy of 97.6% within 106.5 ms from the beginning of the attack. † † cc-license: by 1. Introduction As the 5th generation (5G) of cellular networks scales up to support enhanced mobile broadband, ultra-reliable low-latency connections, and massive machine-type communications, ensuring high availability in the protocol stack control-plane emerges as a major security concern. Both user and control planes are exposed to Denial-of-Service (DoS) attacks that can degrade or interrupt network service ( 11 ; 13 ; 5 ) , but control-plane attacks are more concerning, as they can be carried out by unauthenticated users. Among these attacks, the Radio Resource Control (RRC) signaling storm constitutes a relevant risk for the control-plane. In this attack, an adversarial User Equipment (UE) , which we will refer to as Malicious UE (MUE) throughout the rest of the paper, repeatedly generates RRC Setup Requests with different Radio Network Temporary Identifiers (RNTIs) without sending the RRC Setup Complete message ( 16 ) . In this way, the MUE avoids authenticating to the network but forces the Next Generation Node Base (gNB) to allocate resources for its session (as shown in Fig. 1(b) ). This malicious behavior depletes control-plane capacity and prevents legitimate connections ( 15 ) . The detection of these attacks is challenging because the initial observable behavior mimics genuine high-load events in which the gNB observes bursts of RRC Setup Requests within short time windows. This makes rate-based thresholds ( 21 ) prone to false alarms. Mitigation is even more challenging and far less explored in the literature. Since MUEs never complete the authentication procedure, their temporary identifiers (e.g., RNTI ) change at every new connection attempt. This makes it difficult for the gNB to attribute subsequent connection attempts to specific UEs , and, thus, to implement per- UE blocking policies. Consequently, effective mitigation must rely on device- or transmission-level fingerprints (e.g., Timing Advance (TA) ) ( 6 ) . (a) Normal UE Behavior (b) Malicious UE Behavior Figure 1 . Comparison between Normal UE and MUE Behavior Comparison between normal and MUE behavior We leverage O-RAN as an architectural enabler for addressing the above challenges and design our solution. By exposing standardized, open interfaces toward a Near-Real-Time (near-RT) RAN Intelligent Controller (RIC) , O-RAN provides the observability and control loops needed to collect fine-grained Key Performance Metrics (KPMs) and react rapidly in case of an attack ( 3 ) . Specifically, we introduce StormShield, a near-RT detection and mitigation algorithm for RRC signaling storm attacks that clusters UEs by using multiple radio-level indicators— Received Signal Strength Indicator (RSSI) and TA —and access pattern features (e.g., per-source burst of connection attempts and request-to-completion ratios) to build a robust MUEs fingerprint. The combination of multiple indicators is crucial, as existing defenses—mainly detection-only or volume-based—struggle to separate genuine high load from attacks, and TA alone is insufficient in dense networks where UEs share similar values for this parameter. Differently from prior work, StormShield (i) implements an effective attack detection and mitigation, rejecting connection attempts from UEs flagged as malicious before gNB resources saturate; (ii) discerns high-load vs. attack scenarios; and (iii) remains robust in challenging environments, including multiple simultaneous attackers and mobility. We prototype StormShield on an Over-The-Air (OTA) testbed with programmable 5G protocol stacks (OAI, NVIDIA Aerial) ( 9 ; 10 ) and two different gNB setups. In the first, OpenAirInterface (OAI) drives an USRP X410 Software-defined Radio (SDR) with 8.1 functional split to perform signal transmission and reception. In the second setup, instead, the OAI protocol stack connects to the GPU-accelerated NVIDIA ARC solution, which implements the lower-layer functionalities of the gNB protocol stack, and connects to a commercial Radio Unit (RU) from Foxconn with 7.2 split. We prototype the malicious UE by customizing the RRC state machine in the OAI software UE , and use commercial smartphones as legitimate users. Experimental results in the different setups demonstrate that StormShield is effective in detecting and mitigating signaling storm attacks before the gNB resources are depleted, achieving a 97.6% average accuracy. Our key contributions are the following: • Attack feasibility in realistic 5G setups - We reproduce the RRC signaling storm attack over different OTA testbed setups, including effects such as mobility and multiple concurrent attackers. • Fingerprint-based detection of signaling storms - We design a detection algorithm that combines RRC -layer statistics with spatial fingerprinting. This allows the system to distinguish signaling storm attacks from benign high-load network conditions even in challenging indoor scenarios where distance-based features alone are not sufficient. • Pre-authentication mitigation via fingerprint blocklisting - We introduce a mitigation algorithm that operates before authentication and without relying on stable user identities. It maintains an adaptive block list of malicious fingerprints with aging timeouts that reinforce blocking for persistent attackers, while automatically expiring stale entries to avoid long-term exclusion of benign UEs . We demonstrate the effectiveness of this technique in the presence of multiple attackers, and both static and mobile devices. • O-RAN-based closed control loop implementation - We deploy StormShield in an xApp over the near-RT RIC , enabling closed-loop detection and mitigation without modifying the Radio Access Network (RAN) data plane. We extend the O-RAN E2 data-exposure capabilities with a new security-oriented Service Model (SM) , implemented according to O-RAN use case ( 16 ) , that defines telemetry and control semantics to obtain the KPMs required by StormShield. The