In my first blog in this series, I highlighted some of the main challenges teams face with trying to scale mainframe DevOps. To get past these hurdles, the key is to develop an incremental approach that enables teams to capture value along each step of the journey ...
As enterprise DevOps becomes increasingly mainstream in the year ahead, a flurry of innovations and new improvements will further its progression, fueling more ambitious software development strategies worldwide. Here are a few of our predictions for how enterprise DevOps will likely evolve in 2018 and beyond:
Tim Buntel, VP of Products, XebiaLabs
■ In Cambridge, Mass, you can't throw a VC without hitting a machine learning startup founder. Yet the adoption of machine learning in DevOps has been surprisingly slow. A large part of DevOps is automation. Automation produces data. Frequent automated deployments produce lots of data. And for DevOps teams, there's gold in that data. In 2018, we'll likely see more machine learning techniques applied to optimizing software quality and delivery. Machine learning can help identify ways to improve team efficiency throughout the entire process — from idea to customer value.
T.J. Randall, VP of Customer Success, XebiaLabs
■ In 2018, customers will demand access to information that will help them evaluate the maturity and effectiveness of their DevOps initiatives. This data, along with features for analytics and reporting, will be necessary for proper planning of their continued transformations.
■ Organizations will increase their focus on governance of the entire delivery pipeline, going beyond simply automating deployment to certain environments. This shift in priorities will ensure consistent delivery across the entire path to production.
■ Look for more emphasis on using a single, unified approach to cover diverse technology stacks. One effect of this trend is that it will no longer be acceptable for the CTO to have different development units solving deployment release needs within silos.
Andrew Phillips, VP of DevOps Strategy, XebiaLabs
■ As organizations continue to see the efficiency and ROI gains that come from doing enterprise-wide initiatives, fewer individual teams will have the option to "go it alone."
■ More than ever, companies will expect to see best practices and real-life examples of large enterprises that have implemented DevOps all the way through production. Related to this will be a decreased tolerance (perhaps prematurely) for "Wild West" experimentation and an emphasis on "getting it right first time."
■ Demand for hard numbers and other data that justify the cost and time required for a "DevOps transformation" will rise, as will interest in DevOps platforms that make it easy to understand that data.
Andreas Prins, VP of Product Development, XebiaLabs
■ An increasing number of organizations will discover that software creation is their foundational activity and that they must improve it to survive and thrive. Those who can continually remove barriers in their delivery process will build a competitive advantage.
■ Continuous delivery is not owned by IT, release management, or any one department alone. In 2018, more organizations will come to understand this and will adopt a "dual-mode" approach, which enables everyone in the pipeline to work in an optimal way. Companies that use this method will fast forward their delivery process.
■ In 2018, we'll see a growing understanding of how DevSecOps can solve many security and compliance issues. Organizations that integrate security into their software delivery pipeline will disentangle new energy and capacity in their development departments.