If you’ve read any of my blogs over the last couple of months, you’ve sensed a common theme. At the risk of belaboring the point, I’m writing about it again—the evolution of decentralized IT.
Without intending to, I’ve written a blog series…
- In the first post, I introduced the concept of decentralized IT and described how Hybrid Cloud Management Should Bridge The Worlds Of Centralized And Decentralized IT.
- I expanded on the theme a bit when I wrote about the importance of accurate cost benchmark data when optimizing hybrid and multicloud costs.
- And most recently, I wrote about the coming trend of enterprises filling open Infrastructure and Operations (I&O) leadership roles with people who have zero infrastructure and operations experience.
The through-line of these posts is the evolution of a decentralized IT model in response to exploding public cloud usage and the consumerization of IT services.
As I wrote in the first post, most enterprises have evolved both centralized and decentralized IT models with the decentralized model being the more recent trend. The decentralized model, by virtue of being easily consumed with little to no involvement by centralized IT, created a new set of management challenges as costs grew quickly out of control and no one seemed to be acting as a governor on the throttle. CloudHealth grew to a $100 million business and eventual VMware business unit by bringing order to this chaos. At the same time, data center technology providers like VMware have been adapting and adding new features to their management software to embrace private and hybrid cloud services and help enterprises deliver centralized IT more like a cloud service.
Naturally, delivering data center infrastructure and applications as private/hybrid cloud services creates some of the same challenges as managing public cloud services (monitoring, consumption, chargeback, etc.) and therefore requires some of the management disciplines of the decentralized model. With that in mind, the question going forward is whether these two separates, but overlapping, disciplines will remain separate or will merge.
What prompted this question and this blog post was a conversation I had with a senior-level industry analyst at a recent conference. His opinion was that it wasn’t sustainable over the long-term to have two distinct management disciplines and platforms. They felt it needed to converge at some point. I understand that it’s just one opinion and I know there are others who feel differently but I think it’s an interesting point to explore.
The argument for merging
- IT is best equipped to govern the throttle. For decades, Centralized IT has managed risk and costs, and infrastructure and operations (I&O) teams have a ton of experience creating and enforcing policies and managing governance and compliance. They also know how to work within tight budgets.
- IT has become much more agile. IT has increasingly embraced DevOps and other methods to provide better visibility into end-to-end application delivery processes from the backlog to production. I&O teams are more responsive to developers and users that, in the past, and have more capabilities for delivering IT as a Service.
- Even public clouds are IT so IT should own it. this is partly a pragmatic argument and partly political. The argument is that even public clouds should come under IT’s centralized control. IT teams are the subject matter experts, after all. And politically it maintains centralized IT’s prominent role in an enterprise’s technology purchase decisions and operations. For many, this is the central argument, that cloud is just a natural evolution of IT and centralized IT should own and manage centralized and decentralized IT.
The case for keeping things separate
- The public cloud is too fast for centralized IT. The reason the decentralized IT model emerged and grew so quickly was the need for speed. Traditional IT procurement and operational processes cannot keep up with near-constant launches of public cloud services.
- Different personas for each management model. Cloud management and IT management are used by different personas and organizations. Granted there is overlap in some enterprises, particularly in I&O roles. However, public cloud management spans Finance, Business Operations, engineering, development, and nearly every business unit. The cloud management software needs to be intuitive for a very diverse skillset whereas centralized IT tools are used almost exclusively by skilled IT professionals.
- Too unique to merge. Different usage patterns, different buying centers, CAPEX vs. OPEX, etc., the two models are simply too unique to ever fully merge and be managed centrally. It’s why the two models evolved separately to begin with. It’s why VMware acquired CloudHealth. CloudHealth brought a very unique solution designed for multicloud, decentralized IT management. CloudHealth’s leadership in multicloud management is a perfect complement to VMware’s private and hybrid cloud leadership.
Trying to merge centralized and decentralized IT from a management perspective is like building a duck boat. Sure, it can handle both roads and waterways but the tradeoffs made to traverse both means it doesn’t do either particularly well. The same holds true when merging centralized and decentralized IT. The best approach is to maintain both as unique, yet integrated domains but evolve your organization and roles to make things as seamless as possible. This is why CloudHealth evangelizes the Cloud Center of Excellence (CCoE) as a best practice to bring together the disparate personas responsible for optimizing multicloud spending and usage.
It’s also why 65% of open I&O roles will be filled with people with zero I&O experience. It’s also why I keep seeing frantic ads in my Twitter feed saying, “The role of the CIO will be Unrecognizable in Five years!” Successful companies with hybrid and multicloud strategies will evolve their leadership and their organizations to support both models. There’s no compelling reason to merge them, and it could cause more harm than good.