Agile methodologies for Infrastructure projects?

Being surrounded by agilists, I often find most associating agile methodologies with development projects and waterfall with infrastructure projects! To this I would like to share my thoughts and a recent experience of a large scale infrastructure project for an Energy and utility client. Read on to know more!

To start with, the type of project (software development / infra / network / manufacturing / real estate etc) does not indicate if it suits agile or waterfall method. A project that has more unknowns, where the requirement and the end product is meant to evolve during the project execution would fit well in agile execution. But if the clients are clear on the ask, the team clear with the requirements and you don't expect any major deviation then you would have a choice to execute it in agile/waterfall. Although in such case you may not be reap all agile benefits.

Coming to the infrastructure projects, in most cases infra projects have defined tasks, a well defined config changes, prior similar experiences in place eg : storage migration, infra augmentation on-prem etc. Such projects well fit into waterfall but not all infra projects have defined boundaries, tasks, outcomes or impacts. For instance a project intending to migrate varied and large scale on-prem applications to cloud or a data centre migration project has enough "unknowns" to be qualified for agile.The synergy across various teams (infra/application/network/firewall etc) you achieve when executing such infra projects in agile would be immense.

I have been part of an infra project intending to do Storage migration and HCI/DPS migration across a large data centre. With multiple vendors orchestrated in SAFe agile, this project was a fantastic example. Few of my experiences while on the this project -

1. There were quite a few POCs performed in the project and with agile, we as a team could mark them uncommitted within the first PI until the grey areas were understood. This settles a lot of unrest in business if the project was to run in waterfall
2. The scrum team consisted of resources across client vendors ensured we had their engagement and accountability which is a major challenge in large scale infra projects
3. At times, the steps to be executed for the migrations were fairly similar for most of the servers, so it was easier to estimate the stories than it would have been for a development project.
4. Inspect and adapt helped us improve every program increment. This again was only possible with SAFe agile.We barely see an infrastructure project running in waterfall method take a sigh of relief midway, look back, learn and improve.
5. Delivering incrementally, we categorised the servers based on the applications hosted on them. Business could realise the objectives sooner than they could have if the project was run in waterfall method.?

One can also use Cynefine framework to figure out if the project would benefit from agile or waterfall. This framework categorises the projects into 4 quadrants - Clear, Complicated, Complex or Chaotic. The projects loosely fit into one of these.

• When a project's scope is well understood and the end product is clearly visualised, the project falls in "Clear" quadrant. Eg Support projects
• When there multiple solution options for a well defined end product is present, the project falls in "Complicated" quadrant Eg An automation project with multiple automation tools like Chef, puppet etc or an infrastructure project involving varied teams across vendors and possible impacts on hosted applications.
• Innovation projects fall in "Complex" quadrant as the end product may not be known at the start of the project.
• Incidents fall in "Chaotic" quadrant for obvious reasons??

As the one I mentioned, certain infrastructure projects could be "complicated" and in such scenarios Agile is the best bet!

Hope you got a fair idea on when to choose agile for an infrastructure project. Do share feedback and experiences if you have been part of any infrastructure projects.