Securing Cloud Computing Environment via Optimal Deep Learning-based Intrusion Detection Systems

Cloud computing (CC) has revolutionized the way businesses operate, providing unparalleled scalability and flexibility. However, security concerns loom large with the large quantity of data processed and stored in the cloud. Security is paramount in Cloud Operations to safeguard data, applications, and infrastructure from cyber threats. Intrusion Detection Systems (IDS) have been instrumental in protecting cloud infrastructure by consistently monitoring system and network traffic activities for indications of malicious behavior or unauthorized access. Leveraging advanced anomaly detection techniques and Machine Learning (ML) algorithms, IDS can quickly detect and respond to security risks, helping to reinforce cloud environments against cyber threats. In the digital era, organizations can increase their security posture and ensure the integrity, availability, and confidentiality of their information and services by incorporating strong intrusion detection abilities into CC infrastructures. This study introduces a novel Salp Swarm Algorithm-Based Feature Selection with Deep Learning-Based Intrusion Detection (SSAFS-DLID) method for cloud infrastructure. The proposed method incorporates the SSA for FS, Long Short-Term Memory (LSTM) classification for IDS, and the Adam optimizer for the optimization task. In the context of CC, where security is of great significance, the SSAFS-DLID approach focuses on improving the effectiveness and efficiency of IDSs. The SSA efficiently selects important features from massive datasets, reducing computational complexity and dimensionality while maintaining crucial data. Leveraging LSTM classification, the model can effectively detect anomalies and potential security breaches in cloud infrastructure, offering a strong defence mechanism against different cyberattacks. Furthermore, the application of the Adam optimizer ensures effective convergence and optimization during the process of training. The empirical study highlights the efficacy of the SSAFS-DLID method in accurately detecting intrusions in cloud infrastructures while preserving computational overhead and low false positive rates with 99.71 % accuracy. Overall, the proposed model demonstrates promising potential in reinforcing the security environments of CC system.