Last week we held a webinar on network automation with Forrester Senior Analyst Glenn O’Donnell and US Bank VP Eric L. Cummings. Stay tuned for the link and slides. In the meantime, Eric offered to answer the following questions from the audience via the infrastructure 2.0 blog:

Was there a particular event or series of events that helped you consider automation?

The search for automation was initiated by the sudden growth of our organization in the late 1990′s and early 2000′s. This growth was due to three successive mergers of equal size companies over a 5-year period to form today’s US Bank. It would be nice to say that each company was completely standardized before each merger, but that was not the case. Our organization decided on a path of centralized control and tight standards to ensure a high level of availability to the end user. We had to find an efficient way to ensure consistency and allow for a centrally controlled environment.

How important was automating DNS, DHCP and/or IPAM to your organization?

It was extremely important for us to have one unified method for providing DNS/DHCP and IPAM for our organization. The plans that were created after the last major merger allowed us to re-engineer our DNS/DHCP and network infrastructures. We wanted a cost effective but a highly resilient way to centrally manage these services. Our first attempt at this with a competing vendor allowed us to realize some of the benefits of our centralized management strategy. We were able to provide central control and management with a few staff members, but didn’t fully realize our goals as our company continued to grow.

What were the things, benefits, advantages that your team found most compelling about network automation?

First, we wanted a tool set that would turn repetitive tasks with a degree of human error into simple activities that reduced this potential human error. Second, we wanted to feel that our chosen tool’s manufacturer was a partner in our solution. We wanted to focus more on our network’s design and maintenance versus on how to do things. Our partner would focus on maintaining their automation tools market leadership and help us to meet our needs with future releases. Third, we wanted a tool that would allow us some ability to develop our own scripts for highly complex/tailored activities to significantly reduce the man-hours. One such example is the provisioning of new networks for branches. These activities of entering network, DHCP, DNS, and etc. configurations once took an hour or so to complete. Now it is just a few data entry points and the system does the rest using templates that we created.

What would have happened if you had not automated your network infrastructure?

If we hadn’t automated our network, it would not be as standardized and stable by a significant level. We have been able to ensure proper maintenance of routers and switches. We wouldn’t have been able to maintain a virtually flat FTE level when our environment has been significantly growing. Standardization levels for settings and IOS versions would be poor and our company would have extremely high risk of unintended longer than normal outages.

What are some of the projects that your team has been able to address now that you have automated your core network services?

Currently, my team’s implementation of Infoblox for our IPAM/DNS/DHCP services has allowed us to expand our involvement with 100 percent of the work for our internal environments. We are also looking to expand our services to our external DNS resolution in the coming year. Our team has been able to effectively create a N+1 system with Infoblox that allows us to have extremely reliable and easy-to-implement DR tests. In the past we spent all of our time keeping our heads above water. Now we have time to interact with the network engineers and project teams so that new implementations/designs/etc. consider the day-to-day aspects that our team faces. We have become a partner in the solution versus an automaton.