Learning Puppet 4

Chapter 31. Managing Network Infrastructure with Puppet

While DevOps teams have been building dynamic, push-button node and application deployments, the infrastructure in which they operate has not benefited from the same agility. In many organizations, the DevOps team has to toss a request over the wall to a slow, hand-built infrastructure team. Many other organizations avoid that discrepancy by placing all nodes on a flat network, forsaking the security benefits of functional isolation in order to retain the ability to repurpose nodes dynamically.

Speaking as someone who identified as a network engineer most of my early career, I believe network engineering would be vastly improved by declarative and idempotent implementation. There are hundreds of network planning and management tools—most of which end up being little more than frustrating and limited model makers—to create designs that end up being implemented by hand on the network devices. Infrastructure teams can be slow and inflexible for the same reasons outlined in “Is Puppet DevOps?”—a slower process is necessary to manage and track change, and work around the limitations of the tools available.

Few things are as exciting, and as valuable, as the cultural changes that are achievable when the network infrastructure and security teams can participate in a swift-moving, Agile-based workflow made possible by dynamically declared infrastructure.

Puppet’s ability to configure and audit network devices creates an exciting ...