Four DevOps Lessons to Take Away in 2016
In just 12 months, DevOps has progressed leaps and bounds, moving from the preserve of early adopter and big-name companies to the mainstream enterprise.
To put that idea into perspective, a few years ago, when Amazon, Flickr, and other "unicorns" posted success stories about their DevOps activities, the stories were so amazing that the average enterprise could not and did not relate to them.
For example, if Amazon said it was deploying a new app every 11.6 seconds or Flickr said it was putting 10-plus deployments a day into production, the information was generally greeted with awe and incomprehension, often with thoughts along the line of: "It seems so unlikely. We're in a different industry. To what extent is this actually applicable to us?"
Lesson 1: DevOps Is Doable and It Works
This DevOps theme became more common last year in the enterprise: It isn't easy and it doesn't happen overnight, but it is doable and it works.
First step along the way: Rid your mind of Fear, Uncertainty, and Doubt (FUD) around topics such as regulatory compliance, security, or release quality. Let's get rid of the FUD factor. Let's look at what is actually happening, what competitors have achieved, and what can be learned from their DevOps efforts.
Lesson 2: The Successes of Companies in the Same Vertical Are Compelling
What we've noticed over the past year is that more companies in different verticals publish their DevOps success stories, revealing their real-world struggles and listing improved deployment statistics—statistics that are not "at the unicorn level," but all the more believable and compelling as a result.
For example, Capital One noted last year that it took the company three years to progress from waterfall to a reasonably good Agile position. The company said it was doing 220 applications, one deployment a month—not one every 11.6 seconds—but definitely an enormous improvement over what it was doing.
In a series of presentations about Agile and DevOps, the financial services company mentioned that one of the biggest lessons it learned is that manual testing has to give way to automated testing. This company's experience with DevOps has been informative to other financial services companies seeking to embrace DevOps and looking for a believable role model.
Lesson 3: Cultural Changes Are Key to DevOps Success
The old way of doing things does not work well in an Agile/DevOps experience. It is time to bury the ideas that Dev and Ops inhabit different worlds and that everybody should follow a conveyor-belt approach to logging bugs and fixing problems.
A superior approach, which is being adopted by more enterprises, is that everybody in DevOps is part of a team that shares responsibility for the integrity of the entire software food chain—from creation through testing through operations.
Everybody in the chain has to evolve, to see the big picture, and to stop seeing their responsibilities as simply flagging issues. They must be proactive and ensure they do everything they can to resolve issues. Such thinking has taken root in many enterprises over the past year, becausee companies learn from others in the same verticals and adopt best practices for doing DevOps.
Lesson 4: DevOps Is a Bottom-up Initiative That Needs Top-down Engagement
The impetus to embrace DevOps generally starts with IT people (Dev and/or Ops) who are, of course, the people who will implement the process. However, to take DevOps beyond a small number of teams, IT people have to sell the business benefits of DevOps to senior management or the whole idea will go nowhere.
C-level management makes the big decisions and holds the purse strings. To move your DevOps plan from one app or a small pilot to a second app or the next level, you will need to get C-level approval to make it happen quickly and smoothly, and with the appropriate funding.
Top-down engagement means that DevOps has to be understood and approved by the highest levels of the organization. At the same time, the actual DevOps implementation itself is always going to a bottom-up thing. Developers and operations people understand the nuts and bolts of such concepts as Agile, continuous delivery, and automated testing.
The main takeaway: IT people and C-level people need each other to make DevOps a success.
Mainstream enterprises are doing DevOps big-time, buoyed by the successful implementations of many companies across all verticals. DevOps isn't easy but it is very doable. There's no magic bullet or secret sauce. However, if you overcome the FUD factor, learn from the successes of your competitors, and convince your C-level and tech people to tackle the beast together, you can do it!
About the Author
Andrew Phillips is VP of DevOps Strategy for XebiaLabs, a leading provider of software for Continuous Delivery and DevOps. Andrew is a cloud, service delivery, and automation expert and has been part of the shift to more automated application delivery platforms. He contributes to a number of open source projects, including Apache jclouds, the leading cloud library, and is a co-organizer of the DynamicInfraDays container community events.
*** This article was contributed for exclusive publishing on Developer.com ***