The cloud era in IT is really only getting started, especially when one considers how long the client/server era has lasted, as well as how many organizations are still at the stage of just plotting out their cloud strategies. But through the DevOps movement that it has fueled, this relatively new era is already challenging the notion of organizational silos (i.e. between devtest and ops) that evolved alongside the rise of waterfall development from the 1970s onward.
The changing relationship between software development and infrastructure
Familiarity with infrastructure was essential for developers during much of the client/server and mainframe eras. Knowledge of servers, databases, etc. was necessary for creating code optimized for specific environments and workloads. A "developer" typically possessed the full range of skills now distributed across the programmer and sysadmin roles.
Waterfall lessened the importance of this skill set, however, due to its sequential approach. To its credit, the waterfall methodology may have been ideal for a time when commercial software from any given organization was usually shipped only once or a few times per year on an optical disc. That is, companies could afford the silos and protracted deployment times since everyone else faced similar limitations on agility and distribution, too.
"Web-scale computing requires software to be developed more quickly than ever before."
That's no longer the case, of course, thanks to the cloud and the rise of agile as an alternative to waterfall. Web-scale computing required software not only to be high-quality, but also to be developed more quickly than ever before to ensure its competitiveness. Hyperscale organizations - the Googles and Facebooks of the world - have excelled at the continuous integration and rapid deployment that have become table stakes for Web-scale services.
How can everyone else follow suit? DevOps innovation is one promising possibility, since it ideally enables collaboration across business units, while bringing software development essentially full circle (i.e., to the same programming-infrastructure relationship that existed in the mainframe and client/server eras), as Franklin Morris recently pointed out in a Rackspace blog post. In theory, it enables the agility that many shops need to compete.
DevOps versus traditional ops: What's at stake?
DevOps is primarily a cultural movement aimed at remaking how the entire organization approaches collaboration. To that end, though, it has technological as well as cultural implications. Moreover, its cultural principle of collaboration complements the technical notion of treating infrastructure as code, making it transparent and more easily accessible by all teams than traditional silos - with their programmer bottlenecks - would be.
"What really sets DevOps apart in terms of movement and philosophy can be expressed by the 'infrastructure-as-code' philosophy," observed Shravan Goli, president of Dice, according to CIO. "Traditionally, operations has been an ad hoc type of endeavor - with the capability to track changes and monitor the state of systems being something driven by external processes. DevOps, in general, aims to make those change and state processes more transparent by allowing the traditional concerns of infrastructure, state and security to be shared resources rather than isolated silos of knowledge."
A recent study from Puppet Labs, surveying 9,200 respondents in 110 countries, attempted to delineate DevOps and traditional ops. DevOps turned out to be a time and money saver for many of its adopters around the world:
The last finding could be spun as a positive, since these improvements are increasingly necessary for making the transition from legacy to cloud infrastructure. Overall, DevOps organizations appear set for a stronger future than their counterparts who adhere to traditional ops. More shops in general are warming to DevOps, too: a 2014 RightScale report found 62 percent of respondents planning to adopt DevOps, up from 54 percent the previous year.
With a cloud sandboxing platform like QualiSystems CloudShell, organizations can enable DevOps innovation by giving developers, testers, and other IT consumers the ability to model and provision any type of IT infrastructure. One of the risks in DevOps is the potential for DevOps itself to become another silo, one that is separate from legacy and physical infrastructure on the ground. CloudShell's cloud sandboxing approach that gives end users self-service access to personal replicas of production like infrastructure - spanning data center, network, cloud and converged infrastructure elements, ensures that organizations can make a seamless and effective transition to DevOps practices.
The takeaway: With waterfall receding as the preferred methodology for software development, more organizations are exploring agile as well as DevOps to meet the demands of Web-scale services. Look for interest in and adoption of DevOps to continue to rise in the next few years as teams make the general shift to the cloud era and prioritize collaboration and shared resourcing to ease this transformation.