Dataset for a Systematic Mapping Study of Deep Learning-Based DDoS Detection in SDN and 5G/B5G Networks (2018–2024)

This dataset contains the primary studies used in a systematic mapping study (SMS) on deep learning-based Distributed Denial of Service (DDoS) detection in Software-Defined Networking (SDN), 5G, and beyond-5G (B5G) network environments. The dataset includes 205 peer-reviewed publications published between 2018 and 2024, retrieved from Scopus and IEEE Xplore using a structured search query. Each entry corresponds to a single study and includes bibliographic information (title, authors, year, publication venue etc.), as well as manually curated and derived attributes. In addition to bibliographic metadata, the dataset provides structured annotations extracted from the title, abstract, and manual tags, including: learning_type (Supervised, Unsupervised, Reinforcement, Federated, Not specified; multi-label where applicable) dl_architecture (CNN, LSTM, BiLSTM, GRU, RNN, DNN/MLP, Autoencoder, GAN, Transformer, DBN, Echo State Network, Hybrid, Not specified; multi-label where applicable) network_context (SDN, 5G/6G, IoT/Edge, NFV, Not specified; multi-label where applicable) dataset_used (CICDDoS2019, InSDN, CIC-IDS-2017, CSE-CIC-IDS2018, NSL-KDD, UNSW-NB15, Edge-IIoTset, SDN-SlowRate-DDoS, ToN-IoT, ISCX, CICIoT2023, 5G-NIDD, Other/Custom, Not specified; multi-label where applicable) evaluation_setting (Offline, Testbed-Controller, Real-network, Not specified; multi-label where applicable) The annotations were generated through a semi-automated process combining script-based extraction and manual validation based on titles, abstracts, and manually assigned tags, without full-text analysis. In cases where specific information was not explicitly stated, the corresponding fields were marked as "Not specified". This dataset is intended to support reproducibility and transparency of the associated systematic mapping study, as well as to facilitate further research on deep learning-based DDoS detection in programmable networks.