Hybrid Deep Learning Approaches for Data Security in Cloud Environments

More recent studies have concentrated on methods to improve the security and privacy of cloud operations for the benefit of user data. For example, in order to safeguard sensitive data kept by a cloud-deployed procedure, it is important to guarantee secure transmission of sensitive data across multiple stakeholders. Researchers have investigated deeply into the subject of cloud security concerns, which include data breaches, denial of service, data loss, service rejection, and threats from malicious users. These issues typically stem from problems like multi-tenancy, lack of control over data, and trust issues. Cloud providers are primarily responsible for ensuring the security of the cloud platform. This study proposes a CNN-BiLSTM-AM hybrid model that combines Convolutional Neural Network (CNN), Bidirectional Long Short-Term Memory (Bi-LSTM), and Attention Mechanism (AM). In order to determine how well our proposed hybrid model worked, this study tested it extensively against the traditional CNN and Long Short-Term Memory (LSTM) techniques. An established standard for evaluating intrusion detection systems, the CICIDS 2017 dataset was used in the study. The experimental results demonstrated that the proposed hybrid model outperformed the competition. The proposed hybrid model delivers the highest performance measure of accuracy (99.2%), detection rate (98.807
^%
), and precision (99.59%).