Advantages of Deep Reinforcement Learning in Moving Target Defense for Network Slicing Security

Network slicing networks is crucial for supporting diverse services, requiring stringent security. Traditional AI-based security solutions are reactive, addressing threats post-occurrence. This demo paper proposes integrating Deep Reinforcement Learning (DRL) with MTD to enhance security in network slicing. By dynamically shuffling IP addresses, the network becomes a robust moving target, effectively neutralizing DDoS attacks. Experimental results validate this DRL-MTD framework's effectiveness, paving the way for scalable, proactive defense mechanisms in advanced network architectures.