Cloud is here to stay. While the vast bulk of "cloud" adoption remains firmly in the software as a service (SaaS) category, a non-trivial percentage of cloud usage is in the IaaS category as well. Fifteen percent or so according to IDC's Semiannual IT Cloud Services Tracker (May 2014). While not earthshattering numbers, even 15 percent is significant enough to require attention. In other words — staff.
Staff that's needed to provision, configure and manage resources in "the cloud." The reason there are often dedicated "cloud" people is because managing resources in an infrastructure as a service (IaaS) environment happens within the context of the cloud environment itself. Each cloud has its own unique paradigm, its own vernacular, its own set of APIs and quirky rules for what gets provisioned when and how.
It's a different world within the cloud environment. One that doesn't generally match up with the one inside the data center where resources are being automated and orchestrated using some other system which, inevitably, isn't compatible with the one used out there, in the cloud. While certainly the increased adoption of DevOps in the enterprise is leading to more familiarity with scripting and automation tools and APIs, but these, too, are different than the ones used to manage cloud resources, even when they are managed remotely via an API.
This poses a problem. Not a "get in your bunker, now" kind of problem, but a problem nonetheless. Operational disjunction. Two different environments, two different operational paradigms, two different sets of deployment processes.
I smell a silo coming on, don't you? And we know what silos within IT organizations do, don't we? They cause process obstacles, awkward handoffs, monitoring and measurement difficulties, and generally bad feelings between two groups of people who ought to be working together toward a common set of goals.
The notion that IT must adapt and become adept at managing a hybrid environment — comprised of both cloud and traditional deployments — with alacrity is not a new one, but it is challenging for many organizations. In particular, there is a dearth of knowledge on the subject of just what "skills" are necessary in the first place, though dictates from industry pundits to get those skills abound. Organizations rush to hire those with these nebulous "cloud" skills, who are then grouped together and put to work managing cloud deployments and forming the basis for the next IT silo.
Yet with the exception of SaaS, most cloud deployments are managed via similar methods as those in the enterprises — scripts, templates, APIs and GUIs. What’s different are the interfaces, not necessarily the tools and techniques. Adopting cloud deployment processes that fit a broader, more comprehensive organizational approach — even down to using similar toolsets and frameworks — should be top of mind to reduce the chances of introducing yet another silo into an already ossified set of IT groups.
Heading the potential pitfall of yet another silo in the enterprise off at the pass should be a priority, with those already embracing DevOps in the organization actively pursuing a more collaborative and, one hopes, cohesive approach to integrating cloud and its processes into the fold.
What do you think? Is cloud becoming the new IT silo? If so, how can this silo be prevented in the enterprise?