ML and DL Approaches for DDoS Detection in SDN

EasyChair Preprint 15541

Abstract

Software-Defined Networking (SDN) revolutionizes network management and adaptability by separating the control and data planes. However, its centralized nature exposes it to vulnerabilities, particularly Distributed Denial-of-Service (DDoS) attacks. To address this, Machine Learning (ML) and Deep Learning (DL) techniques have gained attention as effective tools for anomaly detection in SDN environments. This study provides a comprehensive comparison of ML and DL methods for identifying DDoS attacks in SDN. By analyzing different architectures, datasets, and performance metrics, we highlight their respective strengths and weaknesses. Our experiments reveal that DL approaches offer superior accuracy and scalability compared to traditional ML, albeit with increased computational demands.

Keyphrases: Software Define Network, deep learning, machine learning, model, network

@booklet{EasyChair:15541,
  author    = {Maria Kin and James Rajez and Renee Alvez and H Kong and Elvard John and Dela Ahar and Mehmmet Amin},
  title     = {ML and DL Approaches for DDoS Detection in SDN},
  howpublished = {EasyChair Preprint 15541},
  year      = {EasyChair, 2024}}