Smart Jamming Attack and Mitigation on Deep Transfer Reinforcement Learning Enabled Resource Allocation for Network Slicing

TL;DR

This paper investigates a deep reinforcement learning-based jamming attack on network slicing and proposes a DRL-driven mitigation strategy, demonstrating significant improvements in network performance under attack conditions.

Contribution

It introduces a novel DRL-driven attack and a corresponding mitigation model for securing network slicing against intelligent jamming threats.

Findings

DRL-enabled jamming causes 50% throughput degradation

Mitigation improves throughput by 80% under attack

Latency is reduced by 70% with mitigation

Abstract

Network slicing is a pivotal paradigm in wireless networks enabling customized services to users and applications. Yet, intelligent jamming attacks threaten the performance of network slicing. In this paper, we focus on the security aspect of network slicing over a deep transfer reinforcement learning (DTRL) enabled scenario. We first demonstrate how a deep reinforcement learning (DRL)-enabled jamming attack exposes potential risks. In particular, the attacker can intelligently jam resource blocks (RBs) reserved for slices by monitoring transmission signals and perturbing the assigned resources. Then, we propose a DRL-driven mitigation model to mitigate the intelligent attacker. Specifically, the defense mechanism generates interference on unallocated RBs where another antenna is used for transmitting powerful signals. This causes the jammer to consider these RBs as allocated RBs and generate interference for those instead of the allocated RBs. The analysis revealed that the intelligent DRL-enabled jamming attack caused a significant 50% degradation in network throughput and 60% increase in latency in comparison with the no-attack scenario. However, with the implemented mitigation measures, we observed 80% improvement in network throughput and 70% reduction in latency in comparison to the under-attack scenario.