DDoSNet: A Deep Learning Model for detecting Network Attacks in Cloud Computing
Doddi Srilatha, N. Thillaiarasu · 2022 4th International Conference on Inventive Research in Computing Applications (ICIRCA) · 2022
Cloud services are the major target for malicious attacks as more organizations and individuals shift to the cloud. When DDoS assaults pose one of the most significant dangers to Internet users and cloud services, this is especially important. As a result of DDoS assaults, cloud services cannot transmit optimum network infrastructure and sensitive information. Due to the flexibility of cloud computing, DDoS mitigation techniques in the cloud are significantly different from those used in conventional networks. When it comes to DDoS assaults, this article examines their influence on cloud resources and proposes a realistic intrusion detection system (IDS) referred to as DDoSNet. To evaluate models, this research study utilizes the CICIDS 2017 dataset. The intelligence approach of particle swarm optimization (PSO) is applied for feature selection. The optimal features are given input to the deep neural network (DNN) model to classify the given data as malicious or benign. The accuracy rate, recall, precision, and F1-score metrics are considered for model evaluation. The experimental results show an accuracy of 099.81%, precision of 099.77%, recall of 099.89% & F1 score of 099.83%. The presented system outperforms the conventional machine learning models. Hence, this model imparts huge confidence in providing security to cloud services.