DevOps is a buzzword. To understand this popular concept that's transforming how software is built, you need to dig a bit deeper. Tommi Berg, Head of Service Operations at Elisa’s Software Services, does just that.
DevOps is both a problematic buzzword and best thing since sliced bread. The problem with buzzwords is that they mean everything and nothing all at once, and people have different expectations for them. But then again, “DevOps” simultaneously contains excellent values and principles that can be truly beneficial. Let’s start with a definition of DevOps, here is mine:
- Collaboration and open culture
- Automation to accelerate delivery and increase quality
- A software driven platform.
From Platform to Culture
I believe that improvements based on these principles allows for more agility and a faster speed of development. The focus on Elisa’s DevOps journey has been naturally on software development, but I can’t see any reason why these principles couldn’t apply to almost all technology business areas. For it’s common knowledge that all businesses will be software businesses in the future.
I would argue that those three main points could also be visualised with these two triangles. The platform implementation is a largely technical task and usually a huge project: a strategic technical decision. But the actual people impact of just building or buying a new platform is usually small. The existence of the platform alone does not mean it’s getting used.
Next up is automation. Automation can be built on top of different platforms, by utilising the capabilities of the platform. When automation starts to build solutions that gets used, the people impact gets bigger and affects a growing number of people, like software developers. Automation is also an area which will drive more and more actual business benefits of the project as a whole.
Lastly, we have culture transformation, where the implementation isn’t just the technical issues, but also includes how people operate on their own and together. This impacts a large group of people, and isn’t just limited to technical folks. The work needed to achieve this is to coach people and form communities to allow smart people to improve their own ways of working.
There is of course a need for a continuous feedback loop between culture, automation and platform. The design of platform should be driven by culture design, so all technical development needs to be designed with users in the front seat. This is how we at Elisa have designed our DevOps strategy.
AI – the Next Automation of DevOps
Automation is the place where we start to see business benefits and it is clearly a space where we can push for further improvements. I see that currently, automation is mostly developed using regular algorithms, sometimes simple and sometimes very complex. But as those “simple” use cases start to get wrapped up, we have to develop ever more complex algorithms. When they get complex enough, the better way to solve them is to use Machine Learning.
This is why Elisa Viihde Service Architect Jere Nieminen is sharing a real-life story at DEVOPS 2018 on Friday 14 December of how Elisa has used Machine Learning to detect anomalies at Elisa Viihde, and fix them before those issues turn into major incidents.
See you at DEVOPS 2018 here in Helsinki, long may the debate continue!
By Tommi Berg, Head of Service Operations at Elisa’s Software Services
Tommi is an advocate of new information technology, especially cloud and containers. Driving better software development by developing new technology and agile processes.