Culture and resistance to DevOps processes appears to be growing at many organizations where DevOps projects have started going mainstream and involving more enterprise groups.
Cloud automation vendor Quali recently surveyed 1,300 DevOps and IT professionals on some of the key barriers to DevOps adoption and found more respondents identifying culture as a top issue compared to last year. Twenty-two percent of the professionals in Quali’s most recent survey cited stagnant organizational culture as the biggest impediment to DevOps—a substantial 8 percent increase over the 14 percent who said the same thing last year.
Quali interpreted the trend as likely the result of more groups getting pulled into broadening DevOps projects at many organizations. Innovation cultures and processes are being impacted in a broader manner as this is has happened. The trend shows why managing culture has to become a top-down responsibility, the firm said.
“As technologies, processes and tools go mainstream, they touch a larger number of people who may need to change the things they do or how they do it,” said Shashi Kiran, chief marketing officer at Quali. “Eventually, change is never easy. It is actually a good thing that culture is there at the top, which means change is touching a lot of people across the organization, which eventually should be net positive.”
An October 2017 report from analyst firm Forrester Research also suggests that more organizations are at least beginning to recognize the need for a cultural shift. Its research shows organizations are making culture a core focus when implementing DevOps projects. The focus is on shifting the mindset of teams to a culture of experimentation, where failure is used as an opportunity to learn and evolve processes instead of as a reason to add more controls and checks. “These mindset transitions are imperative for successful DevOps, and the market trends are encouraging,” Forrester noted.
Other Barriers to DevOps Adoption
While culture is a top barrier to DevOps adoption, it is not the only one. Quali’s survey showed application complexity and legacy processes to be major factors as well, with 20 percent and 21 percent respectively identifying the two issues as among their top barriers to DevOps adoption.
The survey also found that DevOps teams at many businesses are still heavily dependent on their IT organization for infrastructure and continue to use a ticketing system to request and receive required resources.
Nearly 50 percent, for instance said they had to wait for up to one month to get access to infrastructure resources that they had requested from IT, while the wait was even longer at about 24 percent of the organizations in the Quali survey. Only 27 percent—or barely more than 1 in 4—of the DevOps teams represented in the survey get access to requested infrastructure in less than one day.
The numbers are important because slow access to required infrastructure often could have a big impact on the speed and productivity of DevOps teams, Kiran said.
“DevOps teams don’t want to be bogged down by infrastructure. To a large extent, it needs to be invisible to them,” Kiran said.
However, the fact is that environments need to be set up for DevOps and many times there is an infrastructure dependency on the back end. “This may be someone else’s responsibility, but the DevOps team ends up waiting for infrastructure to be set up and that’s frustrating for them,” he noted.
Predictability is important in DevOps environment as is planning and being able to accommodate dynamic change environments, Kiran said. “Self-service is a preferred option where people are involved in environment buildouts, modeling different components and in testing and deploying them,” he said. One-third of the respondents, for instance, said they were strongly inclined to adopt platforms that enabled self-service automation.
The Quali survey found that the degree to which an organization automates processes has a big impact on its ability to enable a self-service capability. More than two-thirds of the organizations surveyed had some form of automation. Of that, 33 percent had implemented a self-service capability for R&D, development and testing initiatives. Generally, organizations with automation and self-service reported improvements in production and agility compared to organizations that had fewer automation and self-service capabilities.
So, where in the development life cycle is most of the automation happening in DevOps environments? The answer isn’t clear because automation is happening in a somewhat siloed manner, Kiran said. “Continuous integration is a starting point and larger organizations adopt continuous deployment as well, based on their internal methodology.” But, for the moment at least, there is no platform that can navigate creation of DevOps environments, expose them for testing and push them into production—all as a continuous life cycle—with full-on visibility, Kiran said.
Kiran’s take on automation is similar to that reported by Forrester last year, which also found silos of automation continuing to be a barrier to DevOps success at many organizations. Forrester has identified “purposeful end-to-end automation” as foundational to DevOps success, especially at scale.
According to the analyst firm, organizations that have implemented complete automation of the CI/CD pipeline have been able to enable rapid deployment. However, despite progress, plenty of opportunity still exists for automation and for deeper collaboration between infrastructure/operations teams and development teams, Forrester has noted.
Quali also created an infographic illustrating some of the finer points of the survey, which can be accessed here.