DevOps Team Invests in Agile Operations Model

Leading a team of developers working on IT solutions during a transcendental shift has been an eye-opening experience.  Traditional software development has a cadence and regularity that makes integrating new processes fairly straightforward.  It helps that a lot of the Agile ecosystem was crafted for and by software teams.  The status-quo is always comfortable.

These days, I find myself in an environment that is anything but regular.  The sheer landslide of products and solutions that are bubbling up as software permeates the traditional infrastructures of our networks and data centers leads to the technical equivalent of ‘Who’s On First?’.  Pair that with a constant need to customize, integrate and create on top of canned solutions drives project complexity through the roof.  To bring order to the madness requires a new look at practices from industry, engineering, and development.

Today’s IT teams are being caught flat-footed, without the processes or workflows to keep up with the pace of business or technology, but the Agile mantra of ‘release early, release often’ doesn’t apply when you need to guarantee five 9’s.  Is there a middle ground?  Can we speed the IT response while still guaranteeing uptime?  The short answer is ‘Yes’, but it requires you ask another question, ‘How determined are you to change?’.

While delving deeply into how to achieve operational transformation is out of scope for this article, I would like to touch on some highlights from our own experience.  Most importantly, you must fully embrace the concept of team.  There really are no ‘I’s’, there is a unified front and the circle of trust is strong.  The team should share a vision of their process which gives them ownership and responsibility.  They should be equipped and trained on the tooling and solutions they will be employing.  Borrowing practices from Kanban, Scrum, Lean and even XP can help your teams organize, prioritize and focus on work.

Another key component is planning.  IT has always needed planning, but today’s orchestration engines, network controllers and automation solutions require a wider and deeper view into your ecosystem than you would normally see in your CMDB.  With the advent of policy-based computing, software-defined networks and virtual network functions driving service delivery, the levels of abstraction have jumped several orders of magnitude.  Standardization and homogeneous environments are critical to delivering the true value of automation.

Finally, borrowing the toolsets of the software engineer, including code repositories, CI/CD pipelining, test driven development and full dev/test/prod release cycles increases the team’s ability to manage the software-defined ecosystem.  Sound like a lot?  The good news is that you don’t have to consume all of this at once.  You will have to see what works and what doesn’t for your teams in your environment.  Just remember to come back to the main question, ‘How determined are you to change?’