64% of organisations will be using DevOps by the end of this year, according to research from Gartner. And it’s easy to understand why, given the potential of DevOps to vastly increase the frequency and speed of code deployments, while significantly reducing the change failure rate.
Puppet and DORA’s State of DevOps Report 2016 illustrates the scale of benefit that DevOps has the potential to deliver – organisations with high-performing DevOps practices can:
Organisations that are already reaping the huge benefits of this approach include the Metropolitan Police and Ministry of Justice.
It is the emergence of cloud and the consequent increase in consumer demand for continuous product improvement that has thrown the importance of DevOps into sharp relief.
To be responsive to business and consumer needs, IT organisations now need to become capable of faster release cycles – even within legacy environments – while remaining committed to quality and security.
A DevOps approach is now needed, and this requires an evolution in skill sets, culture and ways of working. The traditional methods of Development, Operations and Architecture are no longer appropriate, with silo working causing chronic delays and inefficiencies.
With DevOps, software engineering principles are applied both to software development and IT infrastructure, and specialist representatives of Operations, Architecture and Security are embedded in Development teams.
Key to the approach is to “shift things to the left”, i.e. to deploy to production as early as possible and give early visibility to the Compliance, Security and Architecture teams while ensuring that their integrations work from the outset. In this way, DevOps acts as the “grease in the wheels”, helping to promote trust, communication and collaboration between the different disciplines.
Blockers in driving forward the DevOps approach include legacy environments, vendor capability issues, and governance and compliance impeding the delivery pipeline.
The answer seems to be “myth busting” with real evidence: proving that releasing smaller batches reduces risk and makes bugs easier to identify and solve; showing that deploying as early as possible increases transparency across the board; and demonstrating how DevOps enhances rather than impedes governance and control through its emphasis on standardising approaches, policies and tools.
The author, Paul Dykes, is a consultant at IndigoBlue, an Agile, Lean and DevOps consultancy.