Coming to grips with cloud and making choices
Spoiled for choice
Talking to enterprise IT folks today will almost always lead to a familiar theme when discussing cloud computing and all its interrelated implications. First, cloud is now familiar—’normal’ even. This model of using and consuming IT resources has been around for a decade now—defined as self-service, on-demand, programmatically accessible, elastic, scalable and automated in software. It has made entirely new categories in IT applications and development possible, and has seen the wild growth of public cloud providers as well as the rapid proliferation of private cloud platforms for enterprises. And yet most enterprises today are stuck with the task of grappling with existing infrastructure and applications that aren’t made in that mold, and cloud can drag them down. They face a pressing need to update IT workloads and infrastructure to capitalize on ‘the cloud’ or miss out on the competitive advantages it confers.
There’s often scattershot adoption, opportunistic stabs at transformation or migration to better venues for IT workloads. But just as often, there’s frustration at making IT even more messy, when it seems that a more ideal state should be within easy reach. However, it’s a reality for the current market: 90% of enterprises today use more than one form of cloud environment and service provider, according to independent survey data collected by 451 Research. That spans public, private, hosted and cloud services, and also Software-as-a-Service (SaaS) applications. What’s more, a majority of enterprises fully expect their IT operations will continue to span the multitude of choices and providers out there. There’s no ‘one size fits all’—there’s a choice to be made for every application, and for every part of IT infrastructure
What could go wrong?
One thing that inhibits cloud migration and modernization is the multiple external factors that affect every workload, including things like compliance with regulations, complexity and relative importance to the business, established business processes and the potential cost and disruption involved in an application transformation effort. Sometimes this calculus is simple: unless your enterprise is required to keep data in house thanks to strict regulations, document and data backup should go off-site, where a cloud provider can do an infinitely more secure and cheaper job than your datacenter can. The same could be said for email, or any one of a bunch of similar cases. Sometimes the calculus goes the other way—how much pain and consternation will it cause to adapt and migrate a 30-year-old mainframe workload to the cloud, with all that entails? Thinking this through for one application can range from a chore to a guaranteed head-meets-wall outcome, and when cast over an enterprise portfolio of dozens, hundreds, or thousands of applications, it represents a significant headwind.
One answer is to take a systematic approach that starts with a simple concept: the cloud is mature, and each available venue for any given workload will work as advertised, and the approach to onboarding or consuming any of them (public, private, platform-as-a-service [PaaS], SaaS, etc.) is essentially the same. It’s done via engineering around application programming interfaces (APIs) and the connection to the provider. The details, like API syntax of resource type, will vary, but the process is the same. This simplifies compatibility issues (but doesn’t eliminate them) and means that the choice of best execution venue can be made on more arbitrary factors.
Start by making a rational assessment of current IT operations and assets and a list of pros and cons for each workload in turn, and assign priorities based on business need and feasibility instead of treating cloud migration as a technology refresh. This won’t be orderly, but it can be orchestrated based on practical considerations, including people and process. If it’s more work than it’s worth to convince Bob and Alice in accounting that they can do their work faster in the cloud if they just update their thinking and training, let them be and put effort into something more accessible. As holdouts and tough problems are gradually exposed and left behind, it will get easier and easier to go after those holdouts.
On balance, it takes enterprises about 4-5 years before they report that a majority of workloads (75%) are leveraging some form of cloud environment or providers (451 Research, Voice of the Enterprise survey research). At best it’s a methodical process of reshaping investment, practices and business processes — at worst it’s a morass of failed starts, missed opportunities, poor planning and wasted effort. Enterprises should develop an objective, holistic view of their situation that allows for specific actions to be made with a clear eye as to the cost, the process, and the eventual benefits to finding the right venue for any given workload. There’s no right answer for any given situation, but there are right ways to proceed. It’s okay to have a preference for one provider or vendor, and it’s okay to be as independent and cloud-agnostic as possible. It’s okay to prioritize mass migration, and it’s okay to realize some things are better off in the enterprise datacenter. The important part is to develop ways to go into it with eyes wide open and a core understanding of the options available.
Download the latest Pathfinder Report from 451 Research “Choosing a cloud: how enterprises can find the best location for application workloads” here.