Storage performance and recoverability were separate workstreams
The bank wanted shorter process times and a more manageable infrastructure. Production storage and backup were assessed together, but their implementation followed distinct validation steps.
That separation matters because faster production systems do not establish that recovery will work, just as a functioning backup platform does not remove application-performance constraints. The program addressed both needs without treating one product installation as proof of the other. It connected the modernization goal to two operating responsibilities: serving current workloads and retaining a usable path back when those workloads need restoration.
A pilot became a runbook for production migration
The storage work began with a pilot migration and a detailed runbook. After that pilot, the bank’s own IT team migrated production data in two data centers, accounting for dependencies, resources and the cutover schedule.
The runbook transferred more than a sequence of commands. It turned the pilot’s experience into a method the operating team could apply to the larger environment. That is the important relationship between a trial and a rollout: the trial reduces uncertainty only when its findings influence how the real migration proceeds. The bank’s team remained involved in executing the production change rather than receiving only a finished hardware installation.
Backup moved through a parallel environment
A separate on-site Rubrik proof of concept and technology testing preceded the recovery migration. The new environment was brought up beside the existing one, and datasets moved gradually under the implementation team’s management.
This created a controlled transition between operating states. It also prevented the storage pilot from standing in as the sole evidence for the backup design. The migration account reports no disruption to business operations or customer experience, while describing required downtime as virtually zero. That supports continuity during this implementation without establishing a universal guarantee for every future migration.
The resulting improvement appeared in business processing
The bank reported that some processes moved from days to hours, and others from hours to minutes, including financial reporting and period-end work. The account does not publish a workload-by-workload benchmark or an independently measured recovery-time objective.
Cartennas’s assessment is that the project connected performance improvement to a defensible change process. The storage pilot, reusable runbook, bank-led production migration and separately validated recovery environment each addressed a different source of uncertainty. The reported processing gains mattered after the move, but the migration method mattered while it was happening. Together, those elements explain how the bank changed important infrastructure without making the technology transition itself the customer’s problem.