EXPANDING APPROACHES TO FORMAL VERIFICATION INTEGRATION INTO CI/CD PIPELINES USING MODULAR ARCHITECTURE AND KNOWLEDGE REPOSITORIES

Abstract

The article explores the methodological, architectural, and technological foundations for integrating formal verification into continuous integration and deployment (CI/CD) processes using a modular architecture and centralized knowledge repositories. In the context of the widespread adoption of cloud development, microservice-based systems, and DevSecOps practices, there is a growing need for verification mechanisms capable of integrating seamlessly into fast-paced, dynamic, and automated software delivery pipelines. The proposed concept focuses on the reuse of formalized artifacts – such as specifications, behavior models, verified properties, analysis scenarios, and verification results – by storing them in a centralized knowledge base. This repository supports metadata, semantic indexing, versioning logic, and compatibility mechanisms, enabling efficient reintegration of artifacts across different projects or CI/CD iterations without the need for complete re-analysis. The paper presents a modular system architecture that enables the integration of formalized verification steps into CI/ CD pipelines through built-in adapters for popular systems such as Jenkins, GitLab CI, and GitHub Actions. A mechanism for managing verification checkpoints is proposed, allowing automatic storage and updating of verification artifacts upon every code or requirements change. The use of artificial intelligence (via ML models) enhances the system’s capabilities by enabling automated selection of relevant artifacts, prediction of verification complexity for specific code changes, and recommendation of optimal verification strategies.