The world has reached the cloud tipping point, in which more than half of all enterprises have become cloud first. Clearly, we’ve arrived at the cloud computing era, but what does this mean? Will the future be an Amazon Web Services (AWS)-centric world, or is there room for other public cloud vendors, such as Microsoft Azure or Google Cloud to dominate? What about private cloud? And will on-premises apps eventually go the way of the analog telephone?
Curious, we surveyed 135 cloud professionals at the latest AWS re:Invent show in Las Vegas to find out more. The typical cloud professional we surveyed worked at an organization with 1,500 employees within a wide range of industries, and respondents were split between staff (58 percent) and management (42 percent) roles.
What we discovered is that the future looks a bit chaotic. The idea of a neat, homogenous public cloud future is not likely to happen. The future will be decidedly hybrid.
First, we found that enterprises today have a roughly equal number of on-premises and cloud workloads; by 2020, they expect that to tilt much more heavily toward cloud. No surprise there, but we also learned that fully 11 percent of workloads today are hybrid and that figure will remain about the same at 12 percent in 2020.
What drives enterprises toward a particular app deployment model? The respondents told us that they host apps on-premises for security, cost and compliance reasons, whereas cloud delivers reliability, performance, and flexibility. In other words, the answer to the question of where workloads are hosted is a clear “it depends.”
In terms of which public cloud service is preferred by the professionals we surveyed, AWS is the clear leader: 81 percent are currently using AWS in production, with 16 percent trialing (note: this was an AWS show). But don’t give up on other platforms. Nearly half were using or trialing Azure, and 39 percent were involved with Google Cloud. In fact, a survey by the Cloud Native Computing Foundation shows that Google Cloud is a close second to AWS for container and Kubernetes use cases.
So, the race is not over. In fact, most cloud professionals are using multiple cloud platforms. Why? The top three reasons cited were cost-effectiveness, redundancy, and security. Furthermore, 22 to 24 percent were at least experimenting with Azure Stack or VMware Cloud for AWS, two local cloud offerings.
What about on-premises apps? We asked how long on-premises apps will last, and a third of the respondents told us six more years. One in five say on-premises apps will continue for 10 years or more.
Prepare for the future
So, we are clearly in the era of cloud computing, but this cloudy future will include multiple public cloud platforms, hybrid clouds, some local cloud options, and, at least for the foreseeable future, a mix of on-premises apps. In short, the future looks messy. So, what can enterprises do to better prepare for this hybrid world?
First, IT should perform a tool/ecosystem assessment of all existing solutions for compatibility and ensure that the solutions and tools you choose in the future support hybrid environments.
Second, change your hiring bias to look more for generalists than experts. In the old world, people had tighter boundaries and vertical expertise. For example, a storage admin was an expert in storage, but not so much in networking. In the new world, people act much more as generalists and can cover more IT domains with less required depth.
Third, transition from a CapEx footing to an OpEx footing. Make sure you have the appropriate visibility and control over processes that drive costs. When the environment is hybrid, equipment is no longer a sunk cost. The actual OpEx budget is being used every month, and you need a solution that provides the right visibility for on-premises and cloud to track the cost implications of every engineering decision, not only at the end of the month, but on a daily basis (including alerting, if needed, for exceptionally high consumption) to ensure the tight management of budgets.
And, finally, understand the dependency and topology between components and data. Before engaging in the transition to hybrid, ensure that you understand the topology between application components and the data components of the application and all its users. Make sure that the right components and data components are placed, as projects can fail if your planning does not take this into consideration. Note that having data and other components in different locations (on-prem vs. cloud) may be ok, depending on the type and size of data and the security requirements.
As we fully transition to our cloud-computing future, it is an exciting time. Opportunities abound. But now is also the time to prepare for what is sure to be a hybrid future.