A model-based system for on-premises software-defined infrastructure

Abstract

This dissertation develops and evaluates a novel framework for the adoption of on-premises software-defined infrastructure (SDI) within large, skill-based IT organizations. Focusing on a case study of a major US healthcare provider, the research investigates whether cloud-inspired automation techniques commonly associated with DevOps can deliver meaningful benefits in environments heavily reliant on traditional, on-premises technologies. First, a hybrid simulation approach — integrating System Dynamics and Discrete Event Simulation — depicts both project-based tasks and unscheduled operational work within the case study organization. The findings suggest that automating high-volume or time-critical processes can reduce queuing, shorten response times, and lower error rates by addressing the unique constraints that arise when teams of mixed skill levels must simultaneously manage both project deliverables and incident-driven activities. Subsequently, the dissertation applies model-based systems engineering (MBSE) to guide the systematic design of an on-premises SDI management system. Using the Systems Modeling Language (SysML), a reference architecture is defined that outlines the orchestration, code management, and integrations required to enable a unified, programmable environment across servers, storage, and network resources. This architecture leverages existing tools and hardware investments, providing a cohesive layer through which code-driven automation can be deployed and maintained. Finally, a phased implementation roadmap is proposed in tandem with a quantitative business-case analysis. The recommended approach advocates incremental adoption, beginning with tasks that benefit most from automated provisioning and event-driven response. Taken together, this research offers a practical blueprint for healthcare and similarly structured organizations seeking to modernize their IT environments, enhance operational efficiencies, and harmonize DevOps methodologies with existing on-premises systems and management practices.