Adaptive Attack Mitigation for IoV Flood Attacks

Erol Gelenbe, Mohammed Nasereddin · 2025

Read Adaptive Attack Mitigation for IoV Flood Attacks by Erol Gelenbe, Mohammed Nasereddin (2025) 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-2501-10224.txt.

Introduction

Erol Gelenbe and Mohammed Nasereddin design a defense that keeps vehicle-network gateways working during packet flood attacks, in this 2025 arXiv paper. Gateways for the Internet of Vehicles must spot attacks while holding delays and losses low, yet floods can bury the very detector meant to catch them. The authors first measure how floods degrade a gateway in practice.

Their answer pairs a smart forwarder at the gate with an adaptive mitigation system that wakes once an attack is flagged. The system samples the incoming stream to check whether the flood persists, and drops batches of packets to drain the backlog. A cost balance weighs sampling effort against good packets lost. Tests and theory both show the tuned version blunts attacks well.

Catalog record

CREATORErol Gelenbe, Mohammed Nasereddin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-10224

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Adaptive Attack Mitigation for IoV Flood Attacks Thanks: Manuscript received XX, 2024; revised YY, 2024. Thanks: Copyright (c) 20xx IEEE. Personal use of this material is permitted. However, permission to use this material for any other purposes must be obtained from the IEEE by sending a request to [email protected] Erol Gelenbe, Mohammed Nasereddin Abstract Gateway Servers for the Internet of Vehicles (IoV) must meet stringent Security and Quality of Service (QoS) requirements, including cyberattack protection, low delays and minimal packet loss, to offer secure real-time data exchange for human and vehicle safety and efficient road traffic management. Therefore, it is vital to protect these systems from cyberattacks with adequate Attack Detection (AD) and Mitigation mechanisms. Such attacks often include packet Floods that impair the QoS of the networks and Gateways and even impede the Gateways’ capability to carry out AD. Thus, this paper first evaluates these effects using system measurements during Flood attacks. It then demonstrates how a Smart Quasi-Deterministic Policy Forwarder (SQF) at the entrance of the Gateway can regulate the incoming traffic to ensure that the Gateway supports the AD to operate promptly during an attack. Since Flood attacks create substantial packet backlogs, we propose a novel Adaptive Attack Mitigation (AAM) system that is activated after an attack is detected to dynamically sample the incoming packet stream, determine whether the attack is continuing, and also drop batches of packets at the input to reduce the effects of the attack. The AAM is designed to minimize a cost function that includes the sampling overhead and the cost of lost benign packets. We show experimentally that the Optimum AAM approach is effective in mitigating attacks and present theoretical and experimental results that validate the proposed approach. I Introduction Smart vehicles rely on many onboard and road-side IoT devices, and today’s 30 Billion or more IoT devices [ 1 ] are already subjected to numerous cyberattacks [ 2 , 3 ] which disable systems with huge packet floods [ 4 ] . For instance, an attack in 2017 took down 180,000 servers with an overall 2.54 Terabits per second of traffic [ 5 ] . Thus, it is imperative to develop the understanding and technology that can keep both the IoV and the IoT as safe as possible. Unlike the general IoT which mainly involves stationary, low power [ 6 ] or occasionally mobile devices, the IoV operates mostly with numerous continuously moving nodes, i.e. vehicles, that generate, process, and transmit data in real-time, together with road-side monitoring and relaying systems whose complexity may cause the random misdirection and loss of data [ 7 ] . The safety-critical nature of the IoV, which directly impacts road safety and traffic efficiency, demands stringent requirements on data timeliness, reliability, and very low latency [ 8 , 9 ] , making it particularly vulnerable to Flood attacks, which overload the network’s gateway servers that act as control hubs which manage the flow of data between vehicles and the broader network infrastructure, and severely compromise QoS. Smart vehicles may number two billion by 2035 [ 10 ] , so that supply chains and manufacturing networks will also rely in the future on smart transport [ 11 ] . Since smart vehicles depend on timely and accurate data collection and forwarding, cyberattacks against smart vehicles can cause network delays, data loss and inaccuracy, resulting in road traffic gridlock, driver and passenger frustration, delayed delivery of goods and services, and wrong vehicle traffic control decisions [ 12 ] that can lead to road congestion and impair safety [ 13 ] . Therefore, the present paper proposes an Attack Mitigation approach that operates together with an Attack Detector (AD), that acts at the input of an IoV or IoT Gateway to detect attacks rapidly in real time to optimally mitigate their effects. The AD is based on previous work on highly accurate Machine Learning (ML) attack detection methods [ 14 , 15 , 16 ] which determine whether the incoming traffic deviates from the expected normal behaviour. The novel Adaptive Attack Mitigation (AAM) scheme that we propose, rapidly drops incoming attack packets when the AD produces an alarm and actively samples the incoming packet stream in a manner that minimizes the Gateway overhead and the cost of mistakingly dropping benign packets. I-A Related Work on the Security of the IoV and IoT The literature on cybersecurity for the IoV and the IoT, encompasses a range of studies addressing attack and defense methodologies, proposed solutions, and related challenges. Several security architectures, lightweight protocols, and frameworks for autonomous vehicle communication are discussed in [ 17 ] , emphasizing the integration of advanced security solutions. In [ 18 ] , attack and defense studies are categorized to shed light on methods and to identify security challenges for future exploration. In the realm of autonomous driving, the integration of artificial intelligence (AI) into networks and network control algorithms, raises additional cybersecurity concerns that must be considered [ 19 , 20 ] . Cyberattack detection algorithms may be trained offline with calibrated real or synthetic datasets, or online with real data. Most studies then evaluate the accuracy of the resulting trained ADs with real or synthetic datasets, using the following performance scores that are obtained from the total number of true positives (TP), true negatives (TN), false positives (FP) or false alarms, and false negatives (FN), obtained during testing: \displaystyle Accuracy=\frac{TP+TN}{TP+TN+FP+FN}, \displaystyle True~Positive~Rate~(TPR)=\frac{TP}{TP+FN}, \displaystyle True~Negative~Rate~(TNR)=\frac{TN}{TN+FP}. In some cases, other metrics such as the F1 and F2 scores are also used. The research on ADs, their design [ 21 ] and accuracy [ 22 , 23 ] , often use ML methods that are evaluated off-line [ 24 , 25 , 26 , 27 ] with various datasets, rather than with on-line traffic traces. Such evaluations typically neglect the overload created by attacks on the victim devices or systems, which can paralyze the targeted system with floods of packets, in addition to compromising them with malware. TABLE I: Performance Comparison of Attack Detectors for IoV Networks CyberAttack Detection for the IoV Detection Method Evaluation Strategy Evaluation Metrics: Accuracy \mid \mid TPR \mid \mid TNR “Secure attack detection framework for hierarchical 6G-enabled internet of vehicles” [ 28 ] Hybrid (Federated Learning (FL) + Stackelberg Game) Simulation using the UNSW-NB15 dataset 96\% \mid \mid 95\% \mid \mid 97\% “A Transfer Learning and Optimized CNN Based Intrusion Detection System for Internet of Vehicles” [ 29 ] Deep Learning (Transfer Learning + Optimized CNN) Car-Hacking and CICIDS2017 datasets 99.93\% \mid \mid 99.90\% \mid \mid 99.82\% “An Intrusion Detection System for Connected Vehicles in Smart Cities” [ 30 ] Hybrid (Deep Belief Network (DBN)+ Decision Tree (DT)) Simulation using ns-3 and NSL-KDD datasets with validation through MATLAB simulations 99.43\% \mid \mid 99.04\% \mid \mid 98.47\% “Cybersecurity for Autonomous Vehicles Against Malware Attacks in Smart Cities” [ 31 ] Hybrid (Static + Dynamic Analysis with ML) Evaluated using a custom dataset of 1000 malware downloaded from Virusshare.com and 1000 non-malware applications 96.3\% \mid \mid 95.8\% \mid \mid 97.2\% “Cybersecurity in Automotive: An Intrusion Detection System in Connected Vehicles” [ 32 ] Hybrid (Bayesian Networks + Spatial-Temporal Analysis) Evaluated using simulated datasets generated through the CARLA simulator for various attack scenarios

Read the complete text