Book Review: Model-Driven DevOps

Earlier this month at Cisco Live in Las Vegas, I attended a presentation by Steven Carter and Jason King titled Model-Driven DevOps: The Enterprise Automation Framework You've Been Looking For! They discussed the importance of robustly modeling your network data — a quality very dear to me, as you might expect — and having a clear strategy for automating changes, and shared an example automation architecture for reference. It was a refreshing take on introducing network engineers to DevOps methodologies, and the first time I've seen anyone clearly map software development practices to network automation.

Interested to dive deeper into the material than a 45-minute session would allow, I was pleased to learn that their talk derived from a book of the same name that they co-authored. Coupled with the lingering geeky high from Cisco Live, it made for an excellent read on the flight back home.

The book's content is broken into seven easily digestible chapters, each dedicated to a particular step along a path toward network automation:

• Chapter 1: A Lightbulb Goes Off
• Chapter 2: A Better Way
• Chapter 3: Consumable Infrastructure
• Chapter 4: Infrastructure as Code
• Chapter 5: Continuous Integration/Continuous Deployment
• Chapter 6: Implementation
• Chapter 7: Human Factors

Each chapter begins with a short parable featuring a recurring cast of characters (reminiscent of the Phoenix Project and its successor, The Unicorn Project), followed by discussion of the problems at hand and solutions offered.

What appeals to me most about this book is the way in which DevOps topics and conventions are presented in the context of network automation. The authors understand and directly address the differences between network engineering and software development, taking time to explain how tooling can be adapted to fit the networker's needs.

The ideas of consumable infrastructure and modeling network configuration using standardized language — crucial components for network automation strategy — are also thoroughly explored. The book's final chapter shifts focus to address the human aspect of automation challenges, from organizational culture and politics down to concerns of the individual.

I highly recommend this book as an introduction for anyone looking into network automation. Even if you're already well on your way, it can serve as a reference model and fill in some foundational gaps. Not only is the book itself an excellent read, it comes with an array of hands-on exercises to enhance your study.