"DDoS Attack Duration prediction"
收藏DataCite Commons2026-02-26 更新2026-05-03 收录
下载链接：
https://ieee-dataport.org/documents/ddos-attack-duration-prediction
下载链接
链接失效反馈官方服务：
资源简介：
"This dataset provides comprehensive network DDoS attack traffic and preprocessed temporal sequences for cyberattack duration prediction, derived from CICIoT2023 and augmented with real 5G O-RAN testbed captures. It includes 391 labeled attack sessions across five subcategories (UDP, ICMP, TCP, SYN, and SynonymousIP floods), with 49 original features refined to 21 predictive attributes. Each flow is augmented with a remaining_time target variable, enabling time-series forecasting through chronologically ordered sessions. Attacks are categorized as Short (<600s), Medium (600-1800s), or Long (>1800s). Four federated learning partition configurations (IID, Non-IID by Label, Non-IID by Duration, and Fully Non-IID) support distributed learning research. Five real TCP SYN flood captures (80-120s) from a 5G O-RAN testbed are included for domain adaptation validation."
提供机构：
IEEE DataPort
创建时间：
2026-02-26
