Shift to Self-Service Platforms Highlights State of DevOps in 2020
DevOps is increasingly becoming about the use of internal self-service platforms, according to the 2020 State of DevOps Report.
The 52-page report, released on Nov. 12 and co-authored by DevOps firms Puppet and CircleCI, benefited from the input of 2,400 IT professionals. Among its high-level findings is that 60% of organizations have between two and four internal self-service platforms.
“Anecdotally, this is a big shift from what we were seeing just two years ago, when very few organizations were adopting a platform model,” Alanna Brown, senior director of developer relations at Puppet and author of the State of DevOps report, told ITPro Today. “A lot of this is being driven by Kubernetes and cloud-native architecture adoption, but we also see this model working well for legacy environments.”
In terms of what types of self-service platforms are being made available, Brown said public cloud infrastructure is common at all levels of the DevOps evolution. Companies are providing a standardized self-service interface to both public and private cloud resources that has been hardened for security and compliance, she noted.
“We work with a lot of large enterprises that are moving to an internal platform model and providing self-service provisioning for public and private cloud resources to enable development teams to move at their own pace,” Brown said. “As those teams evolve, they start to offer more self-service capabilities, such as deployment patterns and database provisioning.”
DevOps Taking a Product-Oriented Approach
The report also found that mature DevOps organizations are more likely to take a product-oriented approach to development than less mature ones.
Product-oriented means DevOps teams exhibit a product mindset and practice modern product management techniques, such as gathering requirements, providing a roadmap and facilitating onboarding, Brown said.
It also means that platform teams really have to understand their customers’ needs so they can build the right thing.
“Delivering features and releases is only useful insofar as it meets the customer’s needs,” Brown said.
The Intersection of Change Management and DevOps
The report also looked at the intersection of change management practices and DevOps. Organizations whose employees believe their change management is effective are three times more likely to automate testing and deployment than ones where confidence in change management performance is low.
Overall, the original expectation by the report’s authors, according to Brown, was that change management effectiveness would correlate more strongly with DevOps evolution.
“Change management effectiveness increases as organizations evolve their DevOps practices, but the differences between high, mid and low evolution are less dramatic than we expected, though they are statistically significant,” she said. “We think this is due to the fact that the key practices in the DevOps evolution model revolve around the standardization of tools, technology, configurations, infrastructure and patterns.”
In contrast, Brown said change management is deeply embedded in many business functions beyond software delivery and includes things such as supplier relationships, incident management, risk management, value stream optimization and more.