Modern Cloud Security and Automation: A DevSecOps Approach Leveraging AI/ML and Containerization

Containerization, micro services and Continuous Integration/Continuous Delivery (CI/CD) pipelines are some cloud native development methods that have changed the way software is delivered, but that has also opened up the attacker surface and presented some considerable security risks. Silos and traditional approaches to security can be slow to detect, more expensive to fix and disrupt agile processes. This paper deals with such challenges by suggesting a security conscious, AI/ML enhanced DevSecOps, approach that incorporate security measures across the software development life cycle. The primary task is to build a smart automated and audible CI/CD pipeline for active vulnerability identification and real-time threat reduction with attentive operational functionality. It is a combination of Static Application Security Testing (SAST), Software Composition Analysis (SCA), container image scanning, runtime policy enforcement and Dynamic Application Security Testing (DAST) into a GitOps managed deployment platform. The UNSW-NB15 dataset is used to train AI/ML based anomaly detection models in order to detect malicious behaviors with high accuracy. The experimental assessment shows that the proposed pipeline has accuracy, preciseness, recall, and latency overhead of 97.8, 97.2, 97.5 and 3.8, respectively, and it is more accurate, precise, recalls more and has less latency overhead than five state-of-the-art methods on security and operational measures. These results validate the hypothesis that a fully automated and intelligent DevSecOps pipeline can be employed to significantly enhance the resilience of the cloud security without compromising deployment velocity and scalability.