Hybrid IT in the enterprise has emerged primarily as a response to devops. Enterprises with decades of investments in IT infrastructure, and a natural reluctance to store their data elsewhere, suddenly are forced to contend with a significant portion of their new infrastructure being in the cloud. Agile development needs have outpaced the capabilities of traditional infrastructure, driving demand for infrastructure as a service (IaaS).
This impact of devops on the enterprise—driving the adoption of a cloud-traditional hybrid environment—is the obvious outcome. What is less appreciated is how the formation of hybrid IT can affect devops in turn.
Change is a two-way street
Enterprise requirements for scalability, resilience and security don’t disappear simply because of a desired increase in release velocity. Devops practices must be adapted for the enterprise as much as the opposite is true, because the enterprise can be encumbered by more regulations and possess more to lose than the digital disruptors that pioneered devops methods.
Maintaining a strict “bimodal” separation of traditional and transformational systems is a recipe for inefficiency and potential trouble. Effective hybrid IT should be more than simply two parallel silos of on-premises and cloud-based infrastructure. The policies that govern these environments are based on the needs of a single organization, whereas the approach to devops must conform to common policies, particularly as they relate to resilience and security.
It is ironic that devops, which was so named because it broke down walls between development and operations organizations, has resulted in new barriers between traditional and transformational parts of the IT organization. There is much that each could learn from the other.
Devops across the hybrid IT environment requires an integrated approach
To deliver the promise of devops in the enterprise you must support the entire value chain, removing constraints across many groups, platforms, and processes. This approach of “systems thinking” isn’t compatible with silos, considering data such as accounts, patient records, or ticket purchases are likely stored in distributed systems or even older technology.
Devops in a hybrid IT environment requires an integrated application life cycle management approach that is built for any technology, to accurately identify and measure constraints so that they can be dealt with.
In practice, this means that tools used for different aspects of devops can be used by multiple teams, at least, to share information. For example, to improve the accuracy of planning estimates, development time, test coverage and cycle times, visibility across various pipelines is critical—especially when they share resources such as test engineers.
That may mean repurposing—or integrating—existing tool chains into open source tools favored by those who work on more innovative projects. However, the focus should be kept on supporting the removal of constraints to avoid wasted integration efforts that aren’t necessary.
Devops is for the mainframe too
Historically, the word “legacy” has a negative connotation when discussing technology. But looking at it another way, something that’s stood the test of time is “proven.” For example, mainframe technology on Cobol continues to thrive and is maintained because it works, it’s secure, and it would be cost-prohibitive to replace it. Enterprises that try to implement devops approaches that do not include extending this investment expose themselves to cost overruns and delayed projects that require access to data stored on the mainframe.
Rather than focus on recreating applications built for the mainframe, application, and infrastructure modernization can look to leverage the best parts of devops in the mainframe environment. Culture is often a barrier, but mainframe developers can accelerate mainframe application release cycles by leveraging agile methods such as two-week sprints or linking tool chains with other pipelines across the business.
Mainframe developers can leverage the cloud too, using software to offload development and testing workloads to the cloud, reducing the cost of maintaining application quality. Workloads can also be migrated off the mainframe directly to containers and the cloud when it makes sense to do so.
The emergence of devops (and, by extension, hybrid IT) has disrupted much of the enterprise IT status quo. It’s time that both traditional and transformational communities find ways to work toward the common goal of delivering software and services faster with security and quality built into every step of the way.
This article is published as part of the IDG Contributor Network. Want to Join?