Private clouds are a waste of money and hybrid clouds are mistaken investments.
With no published end-user research other than a conference straw poll, the Gartner cloud myth is that there is just one way to “cloud.” If it’s not a public cloud (and presumably an AWS public cloud) then you are just wasting your time and money drinking the Kool-Aid of legacy vendors.
Missing in the Gartner hyperbole, of course, is any quantitative end-user research of how cloud resources are being used or how businesses value IT and cloud resources. The debate is no longer about which cloud to use: public vs. private vs. hybrid. It’s a false argument that presumes there is only one way to “cloud.” It doesn’t consider how businesses value, use and consume cloud resources.
The more important question is not “which,” but “when”: When public, when private and when hybrid? This is where actual research and some data are needed for context.
So what is really going on in the cloud market?
Today, AWS public cloud represents $8 billion of roughly $480 billion in data center systems and enterprise software spending annually, with over 60% growth for the foreseeable future. While AWS represents 1.7% of IT spend, that number will undoubtedly grow much, much larger. Those are impressive numbers but they don’t tell the whole story. How are businesses actually adopting and using cloud resources?
Fortunately, 451 Research has done the hard research that peels back some of the mystery of how businesses are adopting cloud strategies and why. The “Voice of the Enterprise: Cloud Computing Customer Insight Survey Results and Analysis – 2016” provides significant insight about how business of all sizes are adopting cloud strategies. It is probably the most comprehensive survey of cloud users in the industry. The survey was conducted in Q4 of 2015 and we’ll be following this quarterly to see how the numbers evolve.
First, let’s tackle Gartner’s myth that a "tremendous number of private cloud implementations... have functionally failed."
451 Research questioned broad users of cloud services to see how satisfied they were with the cloud services used by their organization.
On a scale of 1-10, IaaS (public cloud) ranked the highest with a mean of 7.8 while hosted private clouds scored 7.2 and on-premises private cloud scored 7.0. Assuming that a failed private cloud would yield an unsatisfied result, there’s a significant gap between Gartner’s perception of the market and what the end-user research actually shows.
Digging deeper, the survey also identifies what kind of clouds businesses are using today and what is planned in the next six months.
On-premises private clouds are currently used by 43.2% of respondents with that number growing to 59.4% in six months. IaaS public clouds are used by 37.2% and are expected to grow to 57.9%. Finally, hosted private cloud is currently used by 34.3% and is expected to grow to 50.5% in six months.
This a good level set of market adoption of cloud types. But how are these different clouds being used?
Cloud Adoption – On-Premises Private Cloud
Next, the survey dives into how each of those cloud services is being used by each type of cloud.
Just over 50% of the respondents using on-premises private cloud are adopting it for the “broad implementation of production applications” while nearly a quarter are using it for the “initial implementation of production applications.
Cloud Adoption – Hosted Private Cloud
A little more than a third of the respondents using hosted private cloud are adopting it for the “broad implementation of production applications” while a third are using it for the “initial implementation of production applications”.
Cloud Adoption – IaaS Public Cloud
Exactly 27% of the respondents using IaaS public cloud adopted it for the “broad implementation of production applications,” while 35% are using it for the “initial implementation of production applications.” A key distinction here is that more than a third of users are using public clouds for “discovery and evaluation,” “running trials/pilot projects,” or in a “test and development environment.” This is significantly higher than you see in on-premises private clouds… and understandably so. The lower cost of entry and significantly lower long-term commitment with public cloud lends itself very well to these uses.
The Business Case for On-Premises Private Cloud
The biggest difference between the types of cloud that are adopted is the business case for the type of cloud that’s being used. While the beginning of the survey answered the “what” and “how” of cloud usage, it’s the business case that addresses the important question of “why”.
For on-premises private cloud, “improved security,” “hardware cost savings” and “improved availability/uptime” led the business cases while “new functionality,” “software license savings” and “create new/additional revenue for the business” were the least selected business cases.
The Business Case for Hosted Private Cloud
In contrast, for hosted private cloud, “hardware cost savings,” “less to manage internally” and “scalability/for peak demand” were the the leading business cases. Similarly, “new functionality,” “software license savings” and “create new/additional revenue for the business” were the least selected business cases.
The Business Case for IaaS Public Cloud
Finally, the business case for IaaS public cloud reveals that “scalability/for peak demand,” “hardware cost savings,” “speed/time to market” and “less to manage internally” were the leading business cases. Once again though, “new functionality,” “software license savings” and “create new/additional revenue for the business” were the least selected business cases.
More than anything else, the 451 Research survey validates that businesses have already adopted public and private cloud. While not completely satisfied, most of the users are generally satisfied with their cloud adoption, with slightly more users satisfied with public clouds than with private clouds. This is in sharp distinction to Gartner’s claims that a "tremendous number of private cloud implementations … have functionally failed”. Cloud technologies are not fundamentally flawed but they are an appropriately complex set of technologies that can be implemented with the right planning, design and governance.
What is not so surprising from the survey is the business cases that show why the surveyed companies chose the type of cloud they did. Rapid scalability, less hardware costs and less internal IT to manage are hallmarks of public cloud adoption. Public clouds also provide cheap and (sometimes) easy resources to test and develop applications and to run trials.
For private clouds, security is clearly a leading driver, which isn’t a surprise either. Security, compliance and control of data will be drivers for private cloud adoption for some time. While public clouds such as AWS will continue to strengthen and validate their security features, enterprises will still selectively choose which applications they will move to public cloud. Enterprises such as Bank of America and Goldman Sachs have invested in their own data centers as they disrupt their traditional IT environments. Others, including General Electric, are rapidly moving to public cloud. The business case for each is understandable even if their decisions on cloud delivery are different. Again, there is no single way for cloud delivery and even the public providers are starting to offer versions of private cloud. In fact, one of the largest private clouds in the world was built for the CIA by AWS. (Edit/Note that the CIA cloud is better classified as a "Community Cloud").
It’s also interesting to look at the longer view of cloud adoption. In this response, it is interesting to see where businesses think they will be two years from now. Clearly public cloud continues its rapid growth but so does private cloud, albeit at a slower growth rate than public.
Finally, I suspect some pundits and analysts will question any research about clouds if there is not a clear and fixed definition of a cloud. Usually, analysts need rigid definitions for anything they measure so they can compare like-for-like technologies without ambiguity. The problem here is that while public and private clouds both provide application delivery, they do not do it the same way and, as this 451 Research has demonstrated, the business cases are often not the same.
For instance, some cloud purists would argue that elasticity and chargebacks are absolutely critical features when comparing different types of clouds. But what if these are not critical features for application delivery in your enterprise? Does that mean it isn’t a cloud? While an online retailer might require cloud elasticity for seasonal surges in transactions, a manufacturer might have consistent and predictable growth that doesn’t require the cost and planning for elasticity. The same is true for chargebacks. In an ideal world, IT resources would be perfectly aligned to business units and departments to track costs. However, for many businesses, chargebacks are highly political and impractical in a shared services model. While it might help analysts to measure the industry, applying a single, static definition of “cloud” across different delivery models is a red herring. It ultimately depends on the objectives of the business.
In the end, the research shows that there is no single way to adopt cloud technologies and there is no reason to believe this will change any time soon. As cloud technologies mature and as more options become available, businesses will ultimately be guided by their best interests just the same as enterprise IT has evolved for the last 50 years. But most of all, it’s important that we base our understanding of business IT on what businesses are really doing and how they are actually adopting cloud technologies. While there are clearly different opinions of how the market is evolving, it will be interesting to see what Gartner’s cloud research reveals.