Enhancing Network Traffic Flow and Mitigating DDoS Threats in Cloud Computing Environments Through Explicit Feature Interaction-Aware Graph Neural Network

Cloud computing has become the backbone of modern information technology structures, providing on-demand scalability, flexibility and cost-effectiveness. However, the growing dependency on cloud services creates new security challenges, particularly in terms of Distributed Denial-of-Service (DDoS) assaults. These assaults attempt to infect a target system with fake traffic, making it unavailable for actual users. In this paper, Enhancing Network Traffic Flow and Mitigating DDoS Threats in Cloud Computing Environments through Explicit Feature Interaction-Aware Graph Neural Network (ENTF-DDoS-EFIAGNN) is proposed. The methodology begins by performing data collection from the CICDDoS 2019 and UNSW-NB15 dataset, followed by preprocessing with the Multi-Observation Fusion Kalman Filter (MOFKF) used to impute missing values and Synthetic Minority Oversampling Technique (SMOTE) is used for data balancing. The Explicit Feature Interaction-Aware Graph Neural Network (EFIAGNN) is used to detect several DDoS attacks, including DrDoS_MSSQL, DrDoS_DNS, DrDoS_SNMP, DrDoS_LDAP, NetBIOS, Lag, Syn, MSSQL, Portmap, UDP, and LDAP. The EFIAGNN’s detection abilities are further improved by the Adaptive Elite Ant Lion Optimizer(AEALO). To mitigate the effects of a DDoS attack, it involves Boundary Integrated Neural Networks (BINN). The efficiency of the proposed ENTF-DDoS-EFIAGNN is validated using measures like accuracy, precision, average connection delay, computational time, Latency, Scalability, Resource Utilization and dropout rate. The results obtained show that ENTF-DDoS-EFIAGNN surpasses existing methods in regard to accuracy (24.20%, 25.11%, and 26.74%) as well as ROC (23.44%, 24.75%, and 25.64%) compared to techniques like an SDN-assisted defense mechanism for the shrew DDoS attack in cloud computing environment (DDoS-HBA-Bi-LSTM), a multi-point collaborative DDoS defense mechanism for IIoT environment(FQBDDA-CC), and Machine learning-driven DDoS attack detection in VANET cloud environments (ED-DDoS-CC), respectively.