Loading...

International Journal of Data Science and Artificial Intelligence - IJDSAI

DEEP LEARNING BASED LSTM-GAN APPROACH FOR INTRUSION DETECTION IN CLOUD ENVIRONMENT


Cloud computing is a rapidly growing technology paradigm with enormous potential. While cloud computing has many advantages, it also poses new security risks. Cloud computing security vulnerabilities have been identified as the most significant impediment to reaping its many benefits. When sensitive data and business applications are outsourced to a third party, these security concerns become critical. In an existing study, research has found that cloud-based intrusion detection systems (IDS) difficult, more time-consuming, and less secure. The paper proposes Intrusion Detection Systems and fully homomorphic elliptic curve cryptography (IDS-FHECC). LSTM-GAN is a deep learning algorithm that detects intrusion and non-intrusion data. To reduce computing complexity and improve security, FH-ECC is utilized to encrypt the input data. The encrypted data is processed using homomorphic operations such as implicit additions and multiplications. In the process of evaluating the proposed mechanism, different metrics like as accuracy, encryption, and decryption time are utilized. The proposed improves the overall accuracy by 8.5%, 9.8%, and 12.5% better than LOA, AODV, RLWE, and WOA respectively. The encryption time of the proposed method is 21.8%, 35.7%, and 39.2% decreased compared to existing LOA, AODV, RLWE, and WOA methods.