Artificial Intelligence Based Strategies for the Protection of SDN Infrastructures Against DDoS Attacks

Abstract

Distributed denial of service (DDoS) attacks constitute a critical threat to software-defined network (SDN) infrastructures, compromising operational availability due to centralization that makes the controller a single point of failure. This review aims to identify scientific evidence on the effectiveness of artificial intelligence-based strategies for detecting and mitigating DDoS attacks in corporate SDN infrastructures. Method: A systematic review was conducted using PICO methodology in SCOPUS, analyzing 248 bibliographic records until June 2025, applying specific criteria that resulted in 20 selected studies focused on AI strategies and critical corporate SDN infrastructures. Results: AI techniques demonstrated significant superiority achieving accuracies between 97% and 99.81% compared to 75% of traditional methods. Hybrid CNN-LSTM models with optimization algorithms reduced false positives from 15% to 0.24%-2% and improved processing speed by 35%. Four main approaches were identified: ML/DL models, adaptive precision, proactive detection, and false positive reduction. Conclusions: Artificial intelligence establishes a new paradigm in SDN cybersecurity, offering proactive, adaptive, and energy-efficient detection, although it requires greater validation in real production environments to confirm its complete operational applicability.