Recently, Mark Casey, CEO of Apcela shared some insights on establishing an inventory of where your applications are currently hosted, as the first step to improving application performance for branches and remote users. This week, he’ll dig into step two.

Once we know where our current applications reside, both on-premises and in the cloud, we can move on to step 2, and consider where future application deployment will likely occur. The general trends suggest a continued and accelerating shift toward public cloud.

At this week’s Gartner Catalyst Conference in San Diego, Research Director Alan Waite gave an update on the State of Private Cloud 2017. He provided Gartner’s latest forecasts on the migration of workloads away from private on-premises based infrastructure, predicting that 40% of both COTS and custom-built application workloads would move to public cloud IaaS by 2020. The survey further forecast that 33% of COTS, and 38% of custom-built, app workloads will remain on-premises. The remaining 27% of COTS and 22% of custom-built workloads respectively will move to private cloud, typically hyper-converged infrastructure, in third party colocation data centers.

In their latest cloud Market Insight report, Cloud Shift: The Transition of IT Spending From Traditional Systems to Cloud, Gartner forecast that overall cloud spending as a percent of total IT budgets will grow from 16% in 2016 to 24% by 2020. Embedded in this forecast is the accelerating ‘cloud-first’ movement where cloud services are the preferred option for new technology initiatives, particularly application deployments as SaaS. This doesn’t mean that every new application will run in public cloud, but cloud is evaluated as the preferred option. According to Gartner, of the organizations using or planning to use cloud services by YE17, 88% of those surveyed indicated that they have a cloud-first strategy.

Complicating a planning exercise like this are the assumptions and inter-dependencies between application and network teams. For example, decisions to retain applications in the branch are often made based on assumptions about the levels of performance and/or availability of the WAN. In Digital Business and Cloud Demand New WAN ArchitecturesMark Fabbi and Neil Rickard suggest, “A two-way dialogue regarding the balance between the capabilities of the WAN and the needs of the application will result in optimized deployment models that would not otherwise have been considered, such as distributing the front end of the application while consolidating the data or distributing the application to regional hubs.”

Once we have a roadmap for our planned deployment models across on-premises, public cloud, and colocation, we’re ready to take a look at step 3, distributing our security on-services and Internet access points, which we’ll dig into next week.

In the meantime, if you’d like to review any of the Gartner research referenced today, all of these reports are available directly from Gartner, or I’ll be happy to arrange a personal briefing to review them in more detail.