Can IT Infrastructure Management be Agile?

IT Infrastructure Management is generally not exciting stuff. When we do our jobs well, no one notices. The emphasis on stability, reliability and mistake-proofing the environment can lead to operational practices heavy with control, checklists, risk-reducing steps and reviews, reviews, reviews. Essentially, the polar opposite of Agile. While the Agile credo of "fail fast" should not be embraced by an organization which manages infrastructure for a living, can other Agile practices not only be adopted, but actually improve service? I believe they can because IBM has been on this journey for two years and the results speak for themselves.

In 2017, our CIO challenged us to transform our management system to one based on a tightly-aligned but loosely-coupled group of global squads. Each squad was given a specific scope of responsibility, staffed with people who had the skills to perform the required roles, and empowered to deliver. This model was less about efficiency and more about effectiveness. Squads had end-to-end ownership of their outcomes, understood their role, and were able to respond very quickly to requirements and problems.

This was in sharp contrast to the prior model, which relied on a more traditional project management structure. There was a lot of planning and analysis which did not always lead to better outcomes, but did add time and cost.  Our people needed to be organized so that they could be self-directed and more responsive - in a word, more Agile. IBM has some of the best talent in the industry, but if they are burdened with too much process and oversight, they will struggle to become high-performing teams. Everyone understood how the Agile model could benefit application development and increase speed to market, but would it help to improve our ability to deliver a resilient and effective IT infrastructure?

Yes, IT infrastructure management can be Agile.

The journey required culture change and course corrections, and two years in we're seeing remarkable results.  By giving squads the freedom to manage their work and own the outcome, they now use their talent more effectively to solve business problems and continuously improve. Each squad has a single funnel of work overseen by a Product Owner and an Iteration Manager. As is now common in the Application Development world, the infrastructure management squads work on an iteration cycle, pulling work from their backlog based on squad capacity and business priority. New work comes in and is prioritized, scheduled and executed much more quickly than before, resulting in rapid completion of both customer requests and infrastructure management activities. By removing multiple layers of overhead and defining squad scope around the work required to deliver a service, responsibility for results is now directly owned the squads themselves. Any risks they take are their own. If mistakes are made, they fix them and learn. Without the old institutional "command and control" system, squads are free to try new things, make decisions, adjust to technical or business changes quickly and have a direct relationship with their customer. The result is that work is being completed in days instead of months, compliance has been replaced with security, localized and duplicate services have been replaced with global, standard services - and it is all done by engaged, excited employees who know the value of the work they do.

Disclaimer: I work for IBM, but my articles are my own.