Summary
The ability to rapidly deploy is a key element of any devops strategy, and the ability to stage and redeploy hardware is no different, especially in a large institution hosting 230+ networks.
Faced with an Institutional requirement to replace ALL of the 230+ border (demark) switches connecting academic and administrative departments to the University backbone in a cost-effective and timely manner, and the complication of dealing with a seriously high-latency practice to prepare networking hardware for first use, we were able to leverage elements of Python, Subversion, and PostgreSQL to massively scale up and speed up our capacity to stage and field new equipment, improving service levels and by extension providing a key element of internal DR/BC practice.
Although the project was left fallow after the initial replacement was completed, the lessons learned from this project have been very valuable when future projects were planned. This talk will focus generally on the problems at hand, focus on the solution architected and implemented, and end with specific points to remember during similar efforts in the future.