It's that time of year again! I hope everyone has happy holidays. I've been staring into my crystal ball and thinking about what's been happening in 2017 and these are my thoughts on what the big themes in DevOps are going to be in the year ahead:
1) BizIT: Whilst we are working with many organisations who are facing the classic DevOps challenge of getting development and IT operations to work better together (particularly where new digital transformation departments are accelerating agile adoption), there are many that are achieving the target state of OneIT and the friction they are feeling is more about the business' acceptance of what agile and DevOps really mean in terms of their commitment and involvement in the processes and feedback loops. As we move towards working in value streams, the lines will continue to blur between IT and the business and we'll go from align, past integrate and into 'IT is the business' and beyond where there is no separation. See what the CTO at Hiscox, Jonathan Fletcher, had to say in his talk at the DevOps Enterprise Summit in London this year.
2) Incrementality: Those of you that know me well know how I like to mash up the English language and so I just made up another word. The point I'm trying to make though is that in 2017 we moved from talking about transformation to evolution. You can say it's semantics but we learned that repeated, failed transformation efforts have created apathy and fear (we say transformation, you hear redundancies) and so we have found it helpful to visualise DevOps improvements as incremental, evolutionary change. Change happens all the time, right - DevOps is just a catalyst and a word you can translate as the latest best practices in delivering technology driven value outcomes. We'll be doing more of this in 2018 and helping companies think little and often. See this picture of Damon Edwards talking at DevOpsCon in Munich this time last year and you'll see what I mean:
3) The Rise of the Experiment: After DOES this year in London I was lucky enough to attend a DevOps metrics workshop with Nicole Forsgren and Jez Humble. Whilst bemoaning the common #fail of not following up on lessons learned in retrospectives, Jez suggesting writing the actions as experiments. This helps us visualise and get excited about what might happen - or not. We've been using the hypothesis a lot in 2017 - asking participants in assessments if there was one experiment they could do, what would it be and what is the expected outcome, driving learners on our DevOps Foundation course to write their personal action plan as an experiment. We've been surprised though, how hard people find it to write a hypothesis. Yesterday I got this blog post from Karl Scotland (do the puzzle!) - it does a brilliant job of explaining why failure must be embraced and how it really is the flip side of success. Experimentation (and learning) is The Third Way, and hard as it is for people in organisations caught between firefighting the technical debt and attacking the mountains of change to get to here, try we must.
4) More Ops in DevOps: Reflecting on Jon Fletcher's talk in the first section of this post, agile adoption is the key driver for DevOps and it does often feel like it's the IT Operations folks that are scrabbling to keep up. We're getting better at understanding what these means on a day to day basis though for all of the ITSM processes. I've found this post from Rob England very powerful to share with people this year to explain how to get to peer reviewed change (a key indicator of high performing tech teams according to The State of DevOps Reports) and also this one from Jon Hall about the DevOps support model. In 2018 we'll have a new drill down course from the DevOps Institute around ITSM for DevOps which will help organisations see where the DevOps target operating model takes IT Operations.
5) Focus on Value Outcomes: Talking of target operating models, what we are seeing in the organisations of highest maturity is a focus on value streams as alluded to earlier. I do find it interesting that the lean tools Toyota came up with decades ago are still so new in terms of their use in the technology space. I am also fascinated that most times we do a Value Stream Mapping exercise it's the first time that group of people have been in a room together. Here's a TED Talk about toast that one of my favourite customers alerted me to that explains why visual collaboration is so powerful. In 2018 I expect to see more organisations working this way, writing value into user stories and measuring the realisation of that value. We'll have end to end traceability and visibility through the DevOps toolchain and see more people doing great stuff like putting electronic assistants in their marketing people's offices so that they can query them on conversion rates in their apps using the AI in their application performance management tooling.
6) Artificial Intelligence: Talking of AI, there's a lot of it about, isn't there? This one's a bit of a red herring though I think for most of the organisations we are working with. Some are still getting to grips with version control and continuous integration and the basics of automation and whilst I agree that the world is likely to look very different in ten years time, I can't see us doing much with this in 2018. But we'll keep an eye on it and use it ways that are easily accessible as above. I think the emphasis is probably on the toolmakers making it easily consumable.
7) Drop in DevOps Teams: I'm not saying a dip in DevOps, I'm saying less DevOps teams as DevOps becomes the new norm. DevOps teams have steadily been on the rise and the market's not sure it's a good thing. It can stifle the spread of DevOps principles and practices - DevOps is EVERYONE's job. So I'm hoping that as DevOps becomes more prevalent in organisations, we'll see fewer DevOps teams as more people are doing it. I'm also hoping that the term 'DevOps engineer' will fade away - see the tweet from Gene Kim below. One day, we'll all just be engineers. And comb shaped people.
8) Chaos Engineering: Whilst it's anathema to many organisations we work with to purposefully break something (that is so fragile, they are so busy and resource constrained) the arrival of easily consumable 'failure as a service' capabilities will help people focus on safety and help us reframe failure as a positive learning opportunity. I often use the analogy of my father, a pilot, spending a lot of time in the simulator so that when something went wrong in the air it was a non-event as he was prepared to deal with. If you haven't seen Sully yet, do! As people use DevOps to tackle technical debt and inject capacity I think in 2018 we'll find more room to experiment with chaos and resilience engineering and by the end of the year it could be moving from experimentation to standard practice.
9) DevSecOps: 2017 was a big year for DevSecOps and saw the term Rugged DevOps fall away. I spoke in the DevSecOps tracks at AppSecEu in Belfast in May and at the fabulous All Day DevOps event, both thanks to our very great friends at Sonatype. There are mixed feelings in the market about whether DevSecOps is a useful term (we did a webcast exploring that very question) but I do feel that security has been a bit of an afterthought and we need to think more about safety culture as I alluded to in the last section. We're also offering a DevSecOps engineering course in 2018 in response to demand - people really do want to build security into their DevOps pipeline at the earliest opportunity. And we think there will be increasing liability for organisations who don't manage application security well.
10) The Definition of Software: This came up in our first DevOps Leader course last month when one of the participants pointed out that the rise of infrastructure as code is blurring the lines along with cloud. Software has eaten the world. And it's looking pretty good.
Happy 2018 everyone - I hope it's good to you!