Overwhelmed? Floundering? Welcome to the Cloud

IaaS adoption has grown from 2% of companies in 2009 to 76% in 2016

Overwhelmed? Floundering? Welcome to the Cloud

We’ve been tracking the shift to cloud for a long time now, both the precipitous and the halting.

Precipitous, as with SaaS, which went from an oddball kind of solution few companies used 10 years ago to a ubiquitous solution in use at more than 95% of companies 5 years ago – and the backbone of many an organization’s software portfolio.  Halting, as with IaaS and PaaS, which went from an oddball kind of solution 10 years ago to…well, something more than that now, with the majority of companies using IaaS but it still accounting for 10% or less of the overall workload for those companies as of 2016.

IaaS adoption has grown from 2% of companies in 2009 to 76% in 2016

This is a watershed for IaaS and Paas, though, an inflection point.  Penetration is ramping up and soon more than 90% of companies will be using them.  More importantly, the amount of work being done in them is ramping up faster too, and will average close to 50% in just a couple more years.

Why is it taking so long?

Cost is one aspect of the slow rise of IaaS.  It has not been economical for a lot of folks to move to cloud, at least not when they do straightforward like-for-like comparisons of the cost of running a VM themselves versus the cost of doing so in a cloud environment.  Despite years of service cost decreases, folks in this year’s Cloud and Network benchmark are continuing to tell me the same thing those in earlier benchmarks told me: “When we run the numbers we can do it cheaper in house.”

What’s changing there is the growing realization that like-for-like migration is not the way to save money with cloud.  Instead, you have to rearchitect the solution to take advantage of cloud services more fully.  As we discussed here recently, this has a cost – vendor lock-in – but one of the benefits is that it can bring costs down. (We’re exploring this in the current benchmark in more detail.)

Another thing, though, is the problem of adapting the organization to cloud. In some ways this is more intractable than cost. (In fact, some folks are willing to accept a 10% cost hit up front to move a solution to cloud, according to early interviews in the benchmark, as long as they get the performance and ease of adoption and migration they are looking for.)  Not only do staff have to achieve a level of comfort with the model, they have to achieve proficiency in using it as well.  This takes dedicated staff time, a quantity notoriously in short supply to IT staffs, and so stretches over long spans of calendar time as a result.

Even as this slow process is coming to fruition (and benchmark interviewees use phrases like “coming up to speed fast now”) they are facing the other aspects of the problem: adapting processes and organizational structures and staffing models to the cloud. This is where words like “floundering” are cropping up in the interviews. Whether they are able to do something straightforward, like re-train SAN managers to be cloud storage specialists, or pursue a more radical shift, like creating DevOps-style scrums of systems, security, and applications staff to be cloud solutions teams, they are all struggling to find the best path forward.

We developed and validated a cloud maturity model as part of the 2016 benchmark; this year we are refining and extending it to provide deeper and more detailed guidance.  We’ll be talking about it more here, as it progresses and as the results become clear.  If you’d like to be a part of the study (and get to participate in early and deeper discussion of the results) just reach out and let me know (I’m john dot burke at nemertes dot com).