Detection of A Novel Dual Attack in Named Data Networking

Distributed Denial of Service (DDoS) attacks in Named Data Networking (NDN), such as Cache Pollution Attacks (CPA) and Improved Collusive Interest Flooding Attacks (ICIFA), can significantly threaten the NDN network. However, most of the previous research has focused on the detection of a single attack by using Machine Learning algorithms and threshold-based methods, so the previous methods are not efficient enough for detecting dual attacks. This paper first proposes a dual attack by combining I-CIFA and CPA, and extracts the network traffic features, including the number of CacheMisses and the number of PIT entries. After analyzing the severe impact of the dual attack, a detection scheme BO-CatBoost is proposed based on Bayesian optimization and CatBoost. Finally, the experiment results show that the proposed detection scheme is robust in detecting the dual attack.