In nearly every cloud migration initiative, a similar pattern plays out: shared services and infrastructure move first to establish the foundation, then application groups follow, with databases typically migrating alongside the applications that depend on them. And then, somewhere near the end of the project, someone adds a line item: file servers.
Files end up last for reasons that feel reasonable in the moment. Latency concerns, permissioning complexity, hard-coded paths, and performance worries make customers hesitant to put file servers in the cloud. The business is running, the shared drives are humming, and there are bigger battles to fight. They’re just files, and files can wait.
But waiting has consequences that are getting more expensive by the year.
What “just files” really means
Part of what pushes file migration to the back of the queue is an unspoken assumption that it’s low-complexity work. Compared to a database migration or an application refactor, moving files sounds like it should be easy.
The category is broader than the word suggests: documents, PDFs, images, video, presentations, shared drives. According to Gartner, unstructured data of this kind accounts for at least 70% of all enterprise data. It is almost always the largest, messiest, and least-understood dataset in the organization, carrying a decade or more of accumulated permissions, nested folder structures, departmental quirks, and legacy formats that no one has touched in years, often alongside sensitive content that lacks any of the governance controls you’d apply to a structured system.
Permissions in particular are rarely the simple part of the job. As Dan Johnson, Azure Team Lead at SoftwareOne, puts it: “Permission cleanup is often more involved than the data transfer itself. Until you analyze access controls, you won’t fully understand migration scope.”
That kind of complexity has a way of revealing itself at the worst possible moment.
The cost of waiting
When file workloads stay on-premises while the rest of the business moves to the cloud, the friction is immediate and constant. Employees toggle between modern cloud applications and legacy file shares, which fragments the experience and chips away at adoption of the tools the organization just bought. You’ve invested in cloud productivity, but your people are still going back to the old server to find a particular file.
Security exposure compounds over time. Legacy file servers rarely meet modern baselines such as zero-trust architecture, encryption at rest, access analytics, and automated retention. Sensitive content sits on infrastructure that isn’t built to govern it. The longer that’s the case, the harder it becomes to defend the exposure.
The reason organizations tolerate that exposure is often financial rather than strategic. Capital is tied up in existing storage hardware, and the depreciation schedule, lease term, or refresh cycle dictates the timing. As Dan Johnson puts it: “We regularly see customers delay file migration because they still have capital investment tied up in existing storage. Usually there’s a lease or a refresh cycle driving the timing.” Whether the delay is an oversight or a contractual reality, organizations end up paying for cloud and legacy infrastructure at the same time, for as long as the old file servers stay running, which is often longer than anyone planned.
The AI case for a files-first approach
The stakes have shifted. The documents, PDFs, and shared files sitting on those legacy servers are the primary input for the AI tools organizations are now actively building toward. Enterprise search, Copilot, generative AI assistants, and intelligent classification all rely on that data being accessible, governed, and in the right place.
Files siloed on-premises are invisible to these capabilities. Every month they remain there is a month the organization can’t capture the full value from its AI investments. That calculus has shifted as AI has become a board-level priority. An unfinished file migration is something executives now notice.
When lift-and-shift isn’t enough
Not all file workloads are created equal. Some can move to standard cloud storage without much ceremony. Others support analytics platforms, application dependencies, and high-concurrency shared environments, and these need infrastructure built for the performance demands they place on it.
When Azure Files isn’t sufficient for throughput, concurrency, or enterprise-grade resiliency, Azure NetApp Files is frequently the right architectural answer. It’s designed for exactly this category of workload: high-performance, latency-sensitive, and with the reliability that production environments require.
But Azure NetApp Files is not a point-and-click migration. It requires deliberate planning involving a workload and dependency assessment, placement and networking design, controlled data movement, performance validation, and a cutover process structured to protect business continuity. Organizations that treat it like a simple lift-and-shift tend to discover its complexity the hard way.
A different order of operations
The organizations that avoid these problems treat file migration as a first-class workstream, not a cleanup task.
That means starting with an inventory of what exists, understanding dependencies, classifying sensitivity, and identifying the redundant, outdated, or trivial content that shouldn’t make the trip. Most organizations are surprised by how much of their file estate falls into those categories. Getting ahead of it before migration makes everything downstream cleaner and cheaper.
File migration also needs to be aligned to business outcomes such as AI readiness, security modernization, collaboration, and governance, rather than treated as a line item in an infrastructure cost-reduction exercise. That framing matters because it changes who cares. When files are positioned as a foundation for goals an organization’s leadership is already committed to, the organizational will to prioritize them tends to follow.
Start with the workload
File migration seems like the kind of work that gets scheduled after the real decisions are made. It’s the workload most likely to be left on-premises when everything else has already moved. But leaving it behind undermines the cloud investment the organization just made.
Delayed file migration carries costs that go well beyond the expense of running two environments at once. The migration itself doesn’t have to be one of them.
SoftwareOne’s Azure NetApp migration services start with the workload, not the tooling. The assessment phase maps dependencies, performance requirements, and migration scope before any data moves, so organizations understand what they’re working with and what the path forward looks like. From there, execution follows a structured sequence, with validation at every stage.
Author
