In general, I expect to see cloud and DevOps moving fully into production as big enterprises adopt them at scale. While this presents some challenges, it also marks a watershed in the evolution of both cloud and DevOps.
1. Expansion of the as-code paradigm to many more areas
We have been hearing about as-code development for a while in cloud infrastructures and release workflows. DevOps teams that adopted infrastructure as-code can provision their infrastructure repeatedly in the same way. This coming year will see it expand in many more DevOps areas. The as-code way of working enables organizations to benefit from shortened release cycles. These are more predictable and have less delivery risk, boosting business agility, responsiveness and innovation. For 2020, we will see the as-code way of working in areas such as security—security as code—whereby security policies are stored next to code, automatically deployed together with the infrastructure and the system to monitor the performance. Alongside security, as-code is also going to be there for governance and compliance, with the service mesh being a good example. This will bring automation and even shorter release cycles for the business.
2. Landing zone team as a default within enterprises
Organizations are starting to succeed in gaining business benefits from DevOps and cloud at scale. In almost all of the organizations succeeding with these, there is a specific landing zone team (aka platform team), which behaves as a service provider ‘platform provider’ to the DevOps teams. This company platform takes care of the common business concerns, like security, compliance, availability, etc., while the public cloud platform is specifically configured for organizational needs. The landing zone brings consistency and flexibility to the usages of the cloud platform with proper DevOps practices. While there are still struggles with roles and responsibilities between the platform and business, I predict that the success of the landing zone team will become a default for organizations.
3. Multi-hybrid cloud management to cause friction
Organizations are gradually moving to hybrid, multi-cloud environments as more multi-cloud management solutions step into the spotlight, for example, Google Anthos and Microsoft Arc. I expect this to create friction between DevOps teams and the multi-cloud manager. That’s because a cloud management platform contains self-servicing and policy capabilities controlling what can be created on the cloud platforms, the same way DevOps teams manage the cloud with their CI/CD pipelines.
The governance between these two is not yet settled, with many organizations still in the middle of finding a way to handle DevOps at scale and to gain maximum benefits from DevOps and cloud. To this end, the cloud management platform brings confusion in roles and responsibilities. The DevOps mindset of ‘You build it, you run it’ is hard to accomplish when a central system manages the environment and, as this is still a relatively new area, the rules of play have yet to be agreed.
4. Governance sets for containerization
Any new development or tech evolution must be matched with new rules and responsibilities. This is the case with the adoption of container technology that’s part of the new era in DevOps. Running container clusters in production requires some strong governance and compliance processes for managing business functionality. While it is fairly new for teams to configure and package business functionality in a container, this packaging and configuration should be consistent across the complete container cluster. The team managing the container platform is best placed to bring this consistency. The different roles and responsibilities between the container platform team and the business DevOps teams are still confusing, but 2020 should begin to see this confusion easing. Organizations and products will start to put in place consistent practices and techniques for container deployments.
5. AI across all areas of DevOps activity
We are used to artificial intelligence (AI) in so many aspects of our lives, so it should be no surprise that DevOps is the next target. There is already AI infused in testing and operations to bring efficiency to the execution and detection of tests and problems. AI will also find its way to developers to support coding and bring advice on language and architecture practices. There are some challenges with this that I hope to see being addressed in some measure over the course of 2020. For example, as the latest World Quality Report from Capgemini and Sogeti points out, while AI can make testing smarter, there is a need to bring on board (or develop) new skills. These are currently lacking in areas such as AI quality assurance strategists, data scientists, and more.
For a chat about Sogeti’s approach to DevOps, please get in touch.