Detection of Cache Pollution Attack Based on Ensemble Learning in ICN-Based VANET

Yao, Lin; Zheng, Zhaolong; Wang, Xin; Zeng, Yujie; Wu, Guowei

doi:10.1109/tdsc.2022.3196109

Cited by 3 publications

(4 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hybrid Heterogeneous Multi-classifier Ensemble learning, proposed by Lin Yao et al (46), is a novel mechanism that utilizes two key parameters: the request frequency for a specific content from the Content Store (CS) cache and the hit ratio of a content in the CS cache. However, it is important to note that this mechanism has a limitation in that it can only effectively identify highly popular content while struggling to detect low-popularity malicious content.…”

Section: Related Workmentioning

confidence: 99%

Q-ICAN: A Q-learning based cache pollution attack mitigation approach for named data networking

Hidouri,

Touati,

Hadded

et al. 2023

Computer Networks

View full text Add to dashboard Cite

Section: Related Workmentioning

confidence: 99%

Q-ICAN: A Q-learning based cache pollution attack mitigation approach for named data networking

Hidouri,

Touati,

Hadded

et al. 2023

Computer Networks

View full text Add to dashboard Cite

“…The in-network caching of popular content is vulnerable to Cache Pollution Attacks (CPA) for nearby consumers which degrade the content retrieval delay by providing nonpopular content by introducing fake interest packets. Ensemble learning can be used for more accurate predictions to minimize the false ratio of detecting CPA and can be applied to high-speed IoVs [78].…”

Section: Related Workmentioning

confidence: 99%

Mobile Edge-Based Information-Centric Network for Emergency Messages Dissemination in Internet of Vehicles: A Deep Learning Approach

et al. 2023

View full text Add to dashboard Cite

With the rapid advancement of Internet of Things (IoT) communication technologies, the Internet of Vehicles (IoV) has gained significant attention for providing the real-time exchange of emergency traffic information among vehicles and Road Side Units (RSU) to improve ultimate driving experiences and road safety. Information-Centric Networking (ICN) has emerged as a novel networking architecture that shifts the communication model from Internet protocol (IP) based host-centric to content-centric architecture. ICN provides support to push and pull-based messages for efficient content dissemination and retrieval by aiming at content names rather than IP addresses. The Mobile Edge Computing (MEC) paradigm facilitates proximity-based real-time traffic applications and services, reducing the content retrieval latency from the core network without the excessive broadcast overhead. Deep Learning (DL) techniques have been tremendously successful in detecting the severity of real-time traffic data. The integration of DL based ANN model for edge-based ICN-IoV brings real-time traffic prediction, content caching, and forwarding of push-based messages closer to the target area. Furthermore, the deployment of mobile edge servers at critical network positions enhances the availability and responsiveness of the name-based content in the ICN paradigm. In this paper, we propose Mobile Edge-based Emergency Messages Dissemination Scheme (MEMDS) to deliver push-based messages delivery at the event-reported geographical location. We also propose a hybrid DL-based Artificial Neural Network (ANN) and MEMDS model to detect and predict the severity of the safety application under real traces from different cities based on specific parameters. The simulation results demonstrate that the proposed scheme significantly improves the data delivery ratio, average delay, hop count, content retrieval delay, and network overhead than DCN and flooding techniques. Secondly, the proposed hybrid model successfully detects the severity of the request with the highest accuracy, precision, recall, and f1-scores values of 96% than benchmark models using real-time vehicular datasets.INDEX TERMS Artificial neural network, deep learning, Internet of Vehicles, Internet of Things, information-centric networking, mobile edge computing.The associate editor coordinating the review of this manuscript and approving it for publication was Yiming Tang .

show abstract

“…Yao et al 27 propose a detection scheme based on hybrid heterogeneous multiclassifier ensemble learning. The generalization ability of ensemble learning can make very accurate predictions on CPA.…”

Section: Related Workmentioning

confidence: 99%

“…Literature 25,27,28 Adopt machine learning method High complexity and high communication resource cost Zhou et al 28 propose a defense scheme based on deep reinforcement learning against CPA, in which whether a data packet is to be cached is decided by a trained intelligent agent. However, this method is not suitable for large-scale networks because of its high complexity and high communication resource cost.We classify and compare the typical CPA defense methods as shown in Table 1.…”

Section: Approaches Advantages Disadvantagesmentioning

confidence: 99%

An efficient caching security approach for content‐centric mobile networks in internet of things systems

Obaidat

et al. 2023

Security and Privacy

Self Cite

View full text Add to dashboard Cite

With the expansion of internet of things (IoT), the IP-based network architecture has been difficult to support the development of IoT. Content-centric mobile networking (CCMN) models are based on naming the content to get rid of address-space scarcity, caching the content at intermediate nodes to provide efficient data delivery, which can solve the development bottleneck ofIoT. The in-network caching is a key factor to enable practical deployments of CCMN. And it is also subject to serious security threats of cache pollution attacks (CPA), which can tamper the distribution of content and reduce the advantages of built-in cache. In addition, the identity and trust mechanism of mobile devices are also important factors hindering the deployment of CCMN. However, existing caching security mechanisms cannot be applied directly, because the features of the built-in cache, including ubiquity, mobility, and heterogeneity, bring new challenges on designing the caching security mechanism against CPA. In this article, we propose a lightweight CPA detection and defense approach for CCMN. We first model the attack and defense process of CPA as a dynamic game, which can optimize the performance of the attack detection. Then, we adopt popularity analysis to detect CPA, and accurately estimate the current network and attack state by a logical regression algorithm. Our design can also intelligently maintain the identity and trust mechanism.Finally, we design a punishment mechanism to deal with the attackers. Extensive simulations demonstrate that our approach can detect CPA efficiently, mitigate the impact of CPA effectively. Our game model can reduce the impact of locality-disruption attack and false-locality attack by 60% and 30%, respectively, compared with the state-of-the-art methods. When the game reaches Nash equilibrium, our model can reduce the impact of CPA on cache hit ratio by 55% on average.

show abstract

Detection of Cache Pollution Attack Based on Ensemble Learning in ICN-Based VANET

Cited by 3 publications

References 28 publications

Q-ICAN: A Q-learning based cache pollution attack mitigation approach for named data networking

Q-ICAN: A Q-learning based cache pollution attack mitigation approach for named data networking

Mobile Edge-Based Information-Centric Network for Emergency Messages Dissemination in Internet of Vehicles: A Deep Learning Approach

An efficient caching security approach for content‐centric mobile networks in internet of things systems

Contact Info

Product

Resources

About