Neural network-based distributed denial of service (DDoS) attack detection in smart home networks

Due to various limitations, such as limited power supply, the lack of storage capability and processing power, Internet of Things-based smart home networks have become vulnerable to various cyber-security attacks including Distributed Denial of Service (DDoS) attacks. These attacks are a malicious attempt to exhaust and overwhelm the target system resources, which has significant impact on the operation of smart home net- works. This paper proposes a novel, efficient and lightweight DDoS attack detection scheme in smart home networks, which employs artificial neural networks (ANN) to classify smart home networks traffic into DDoS attacks or normal traffic. The proposed solution is evaluated on four datasets, namely, IoT-23, DS2OS, NUSW-NB15GT and CICDDOS2019. Experiments were conducted on two types of ANN models, i.e., Multilayered Perceptron (MLP) and Long-Short-Term Memory (LSTM), which achieved 99.78% and 99.98% accuracy, respectively.