Sue Odle is CEO of StorMagic, where she helps organizations simplify on-site virtualization for distributed IT environments.
VMware licensing changes have pushed many retailers to transform their virtualization strategy. For organizations with hundreds or thousands of locations, the challenge is managing the platform migration without disrupting store operations, increasing support requirements or implementing complex configurations that become difficult to manage over time.
Retail infrastructure is highly distributed. Applications supporting sales transactions, inventory, security and other daily functions often run locally without onsite IT support, which means a VMware migration has to account for far more than the technical replacement of one platform with another.
That pressure to achieve optimal infrastructure performance and resiliency is growing. A study found that “only 39% of retailers [are] satisfied with their edge computing support,” even as “AI workloads, from near real-time video analytics for loss prevention to predictive maintenance, move closer to the store level.”
Retailers at scale manage variables across a fleet as hardware configurations evolve through refresh cycles and along innovation paths toward the store of the future. Some elements must remain consistent: deployment methods, unified management and orchestration, and support programs during and after the migration is complete. Retailers need flexible solutions that can account for genuine differences between locations in the current and future operating environment, while still using a migration model that central IT can repeat across the estate. These are nonnegotiables.
Start with the estate.
Retailers should begin by understanding what is already deployed across their locations. That includes workloads that must run locally without dependency on the cloud, what hardware is in place and whether it can be leveraged through the migration, and latency and bandwidth considerations by location to understand where connectivity may affect the migration process and ongoing performance.
Understanding the current state enables IT teams to select the right VMware replacement platform that can be deployed in a repeatable model while accounting for location differences in hardware and connectivity.
Hardware decisions need their own timeline.
A virtualization migration doesn’t require a hardware refresh. Retailers should assess whether existing servers can continue to support the workloads that need to run locally, particularly at locations where hardware still has useful life remaining.
Replacing software and hardware at the same time increases cost and adds more moving parts to each deployment. Separating those decisions can give IT leaders greater control over capital planning and allows hardware to be refreshed according to condition, workload requirements and lifecycle plans.
Remote deployment is a baseline capability.
Most retailers cannot put a skilled engineer in every store during a migration, so the deployment process needs to reflect the people who will actually be available onsite.
IT teams should plan to complete as much of the migration as possible centrally. If a hardware refresh is part of the migration, image and configure systems before they reach the store. Plan for minimal intervention by store staff during installation. The onsite role should be limited to basic physical tasks while the technical work is handled remotely.
A deployment that depends heavily on local troubleshooting becomes difficult to scale. Travel costs increase, maintenance windows become harder to coordinate and central IT spends more time managing individual exceptions. The same principle applies after migration, when teams need visibility into what is running at each location, which systems have migrated successfully and where support is required.
Protect the workloads closest to store operations.
Cloud adoption has changed where many retail systems run, but local workloads will always play a critical role in day-to-day store operations. Connectivity is not guaranteed and varies by location. This reality is driving growth in edge computing because the cost of downtime is enormous.
Workload dependency is an important part of the migration plan. A point-of-sale application carries a different risk than a system that can tolerate several hours of downtime. Inventory, security or other locally dependent applications may also require shorter maintenance windows or stronger rollback procedures.
Research from the University of Illinois shows how quickly technology disruption can reach the business. Researchers studying a two-hour mobile app outage at a large retailer found a 7% decrease in purchases among affected customers and $1.08 million in lost revenue in the short term. The outage also reduced shopping in physical stores, showing the connection between technology availability and the broader customer experience.
Retailers should identify those dependencies before rollout and use them to determine the order in which workloads and locations are migrated. Starting with a representative group of stores gives IT teams the opportunity to test deployment time, workload availability, remote management and support requirements under real operating conditions. Those findings can then be used to improve the process before a much larger portion of the footprint is moved.
Consistency matters more as the estate grows.
The migration method should be consistent enough that central IT knows what to expect at every stage. Sites should be assessed using common criteria, similar locations should follow the same deployment pattern and validation procedures should be clearly defined.
This gives support teams a controlled migration playbook. When engineers know the dependencies of a store, or group of stores, they can address issues more quickly, mitigating the risk of extended downtime.
The store should feel as little of the migration as possible.
Retail IT teams will naturally track technical milestones such as sites completed, virtual machines migrated and workloads validated and restored to normal operations. A successful migration depends on understanding the estate, separating software and hardware planning where possible, designing for remote execution and protecting the workloads that stores depend on.
VMware licensing changes have given many retailers a reason to review their virtualization strategy. They also create an opportunity to improve the way distributed infrastructure is deployed and supported across the business.
Moving to another platform while operating a highly distributed estate in diverse connectivity settings is a significant challenge. For retailers with hundreds or thousands of locations, simplifying the approach so that it can be repeated and automated across the estate can help ensure a successful migration and ongoing management capable of withstanding future change and supporting the store of the future.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
