Apr 21, 2017 Cloud Optimists: Irrational Exuberance, Cloud Edition
Based on our 2016 benchmark, some folks are overly optimistic about how much of their workload will be migrated to public cloud by the end of 2018.
Cloud Optimists: The Glass Is Half Empty
If all you knew about cloud was what you read in any number of breathless articles about it, even the super-breathless ones about multicloud and hybrid cloud, you’d think data centers had been nearly emptied out now that nearly everyone has moved nearly everything has moved to the cloud. Of course, this is not how things really are.
In our cloud benchmark in 2016, we may have seen migration to cloud hitting a long awaited inflection point, going from high growth on very low percentages of workload in cloud to high growth on not so tiny percentages and rapidly rising overall share. (Watch the median line on the chart below.) It depends completely on whether folks were being irrationally exuberant in predicting their own cloud futures, or realistic. And on how this year’s data looks.
Last year, folks were, overall, projecting a doubling of workload in cloud by this time this year. They were looking at the total of what they had to convert and figuring on moving through it quickly given some newfound confidence. The queue, figuratively speaking, was half empty–and emptying quickly.
Cloud Realists: The Glass Is No Longer Full
This year, I am mostly hearing more caution. People are not as far along as they’d expected to be and are not moving as quickly as they thought they would.
Why? Reasons vary, but reason number one is governance. They are not moving more workloads into the cloud more quickly because they don’t know what they are doing (lacking a solid strategy and roadmap) or they don’t know how to organize themselves to do it.
Reason number two appears to be lack of resources. Some lack the financial resources to invest up front in making transitions they expect to save them money in the long run. Others lack the staff resources they need to sort out issues of security, integration, implementation, or migration, and find it is hard for them to bring their existing staff up to speed when they are already overloaded, and just as hard to hire in someone with the skills ready to hand.
And yet…even among the folks who have no strategy, no roadmap, no dedicated staff, and just the littlest toes dipped into the pool – just fractions of a percent of work moved to cloud – some are still boldly predicting that they will move 10%, 20%, 30% to the cloud in the next 18 months. The attractions–agility, resilience, focus, even savings (where savings are possible)–combine with the relentless pressure of leadership to leverage cloud to generate this apparent blithe optimism. SaaS is easy and successful, and moving a few low-impact, low-risk workloads to IaaS is easy and usually successful, and this seems to create the perception that everything else will be too.
But everything else is harder and takes longer than expected. Lifting and shifting VMs is in some ways the easiest approach, but even it is not trivial for complex environments, and it also appears to be prone to increasing rather than decreasing costs. Rearchitecting solutions is not as simple, and takes even longer (but is more likely to yield savings).
If we have any solid guidance from past experience it can probably be found in the server virtualization space, where, after the first 30% or so, every added 10% of workloads virtualized was significantly harder than the previous 10% as the environments being moved got steadily more complex.
I look forward to getting the numbers out of this year’s data and updating the curves above, to see the impact of reality on the cloud optimists’ predictions.