Industry shows dedicated to a single topic results in an inspiring concentration of experts and an exceptional opportunity to learn more about the subject. DevOps Enterprise Summit (DOES) London 2017 delivered on both fronts.
A diverse group of attendees visited our booth. Their job titles included the phrases ‘Transformation Services’, ‘Process Optimization’, ‘DevOps Practitioner’, and ‘Agile CoE’. Very different descriptors, a single denominator – they all wanted to discuss their challenges as they begin to embrace DevOps culture and practices.
It’s understandable. As Mark Levy, Micro Focus Director of Product Marketing and Strategy points out, each enterprise has its own unique DNA, full of complexity, sophistication and chaos. Companies are trying to understand how to adapt their current environment to deliver innovation faster.
Experienced teams who have been continually improving their DevOps practices for years presented many informative sessions throughout the week. Given the inherent independence of DevOps, for these veteran teams the biggest challenge is often how to scale.
Companies attempting to establish transformational DevOps principles are facing challenges on many fronts as they evaluate how to reduce waste and inefficiencies in their application deployment pipelines. With the end goal of establishing transformational DevOps principles, imagine trying to overlay operational and cultural changes when you have a mix of Agile and waterfall processes in place, or a range application environments that span from tightly-coupled to loosely-coupled. Most likely different teams are using a non-integrated, but well-established set of tools.
Automation is often the first step for companies beginning their DevOps journey, to reduce the human error of manual processes, and leverage toolset integrations to speed the feedback loop for iterative methodologies. Dev teams must move fast without breaking things and I spoke with several companies who were first automating code deployments and establishing version control, before moving on to test automation. ‘Move fast without breaking things’ is a common mantra for delivery teams.
Application validation within the context of rapid releases and a complex infrastructure stack (what we fondly called a ‘house of cards’ from my days in the server industry) often introduces schedule and quality risks when manual processes are still the standard. Automated testing and test management should be integrated into Continuous Delivery workflows so functional and performance test results are quickly available to dev teams.
The ability to schedule and control the state of test environments leads to pipeline optimization not only for preproduction environments but also provides a vital link for the handoff to release into production. Today’s applications are deployed across multiple infrastructures (physical, virtual, container, and cloud). Orchestrating complex deployments ensures that components tested together are released together.
DevOps at scale
A large telecom company presented a great DOES session, and outlined their experience implementing Agile and continuous development practices over the past few years. Having determined that software engineering was one of the company’s core competencies, they committed early on to the cultural and organization changes required. Because they viewed organization hierarchies as a ‘pitfall’, with too many roles and too much structure to keep pace with rapid innovation, development teams were given the independence they needed.
They discovered this built-in independence can be an issue when scaling however, and they adopted the Spotify organization hierarchy of Tribes and Squads. Rather than being a barrier, the dependencies between teams are identified, and Squads work regularly to communicate and collaborate.
Guardrails – keeping you on the path to DevOps
The DOES show was promoting many worthy books written by industry researchers and leaders. In our booth, we gave out Gary Gruver’s book Starting and Scaling DevOps in the Enterprise.
In the book, he states “…the benefit of having an architecture that enables small teams to work independently is that they can learn, adapt, and respond more quickly than large, complex, tightly coupled systems.” Independence reinforces ownership for project success, and extends all the way out to working code. This is fundamental to breaking down communication and process silos between different internal organizations.
Gruver emphasizes the need for ‘guardrails’ as a method to enable scale. Guardrails are essentially guidelines that provide a clear understanding of where teams can be independent, and where they need to leverage the same tools and processes to establish commonality across other teams.
One of the most interesting statements I heard was from a company who had transformed their delivery process so drastically, that they no longer saw themselves as software development, but as a manufacturing line.
The visualization was one of efficiency, stability, repeatability. The metaphor also solidified for me how much of the DevOps transformation is being driven from the inside out; dev teams are pressured to innovate faster, and it is this desire to break down barriers, whether cultural or process driven, that fuels a sense of urgency and acceptance across the organization.
Adopting and scaling DevOps at the enterprise level is a multi-year transformational project, and organizations that commit resources and do the hard work see the results. Micro Focus understands the challenges of implementing DevOps for the enterprise and has the experience and credentials to help you on your journey. Check out our dedicated page here.