John Purcell, Chief Product Officer at DoiT.
Upon learning a newspaper had mistakenly published his obituary, a very alive Mark Twain sent a cable to the U.S. press saying, “The report of my death was an exaggeration.” This is how I feel about the public cloud when people talk about workload repatriation—the shift away from the cloud back to on-premise infrastructure and data centers.
The cloud is the beating heart of modern business, circulating data, opening up collaboration and providing access to critical tools and applications throughout an enterprise and across the world at a pace never seen before in the history of IT. It might not be perfect for every company or use case, but the advantages of the cloud are powerful: infrastructure when you need it at the scale you require; the ability to take advantage of a cloud service provider’s (CSP) rapid technology innovation; and the reduction of IT management costs.
Yet, while most everyone agrees about the benefits of building a business in the cloud, some feel pressure to shift back on-prem once established. They believe running a business predominantly in the cloud drags market capitalization over time. They’re also unsure if it’s economical to run a business in the cloud once a certain scale has been reached, and if the cost to maintain that point would be better served on-premises.
Citing high costs alone as a driver to repatriate is a lazy argument and symptomatic of a more significant issue—a lack of sensible governance, management foresight and proactive controls. The good news is that you can better manage your cloud and help pin down costs.
Tooling Around
Repatriation entails moving apps, data and services out of public clouds and back into data centers, on-premises, private or hybrid setups. Companies with data that constantly move in and out of clouds might think repatriation is a great way to avoid ingress and egress fees. In reality, getting data out of the cloud is more expensive than putting it in.
For example, it costs nothing to add data into Azure, but to get it out, the first 50 terabytes cost more than 8 cents per gigabyte. Not only are there significant charges for moving data from the cloud, new hardware is also required.
When it comes to building products, the prep costs to get on-premises technology in place to start the build are astronomical compared to the off-the-shelf, scalable options CSPs offer. A CSP’s suite of tools overcomes pre-build obstacles and accelerates time from brainstorming to development. Hence, the level of flexibility and speed with which you can go from idea to realized product in the public cloud eclipses what you can do on-premises.
There have been waves of disruptive change in IT from mainframes to local desktops. A pattern has emerged: Innovative technology drives disruption and growth, followed by a wave of management tooling to support it. We’re not even two decades into the public cloud journey—yet established companies are still adopting it for the first time today, and so how to effectively and efficiently manage it is still evolving. Even so, tools exist that can help an enterprise make good cloud choices and especially get control of their costs. Independent vendors have made significant progress in developing software and services, some combining technology and expertise for solutions that can help optimize and manage cloud resources. By repatriating, some advantages are gained, but businesses may lose access to the tools and tech available in the cloud today and the tools developed tomorrow as well.
Start Early, Scale Efficiently
The key to responsible cloud management is deploying cloud technology and management processes as early as possible in your company’s growth. Adding management controls back into a scaled environment is difficult, and outside tooling alone won’t get you there: Processes, people and culture are also critical to success. Those elements require a shift in an organization’s culture so employees feel responsible for their own cost efficiency. That way, the following 50 people hired will adopt this same culture of accountability, avoiding any talk of repatriation once you reach scale.
Underpinning it all must be an intelligent FinOps platform, accompanied by initial support on using its tools to examine infrastructure for places where money is being lost or wasted. Common examples include provisioning a virtual machine with 32 cores of compute, when 16 will more than suffice. The platform should also enable automation, such as turning off idle resources and rightsizing consumption. Some solutions apply sophisticated machine learning algorithms for capabilities from advanced forecasting to surfacing anomalies. The emergence and rapid acceptance of GenAI techniques unlock entirely new ways to think about management.
FinOps solutions that classify and enrich cloud billing data make it easier to manage and predict spending and work with users. Be sure cloud costs are indexed to target business metrics (e.g., revenue) in a predictable, desirable way. That way, you can link cloud consumption to your business’s performance and better protect margins. These ideas in the world of IT and corporate governance have been around for decades. There are no practical reasons for not applying modern implementations in cloud ecosystems.
Moving Left Is Right
Engineers are rarely cost-sensitive in their day-to-day work—they’re often tasked with delivering differentiated functionality on a shrinking timeline. So, taking action when a cost issue has been identified isn’t a priority and is often classified as a distraction! Make sure those cultural efforts to instill awareness, responsibility and ownership are heard and followed by these professionals.
Obviously, the death of the cloud is a tremendous exaggeration. Repatriating works for some companies, but not all. Pinning down costs in the cloud isn’t really so tricky these days. Taking the time to learn how will not only create immediate savings but will also support and drive your future business growth in a viable, healthy and exciting way.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
