Zero-Downtime Infrastructure Modernization for a Regional Healthcare Provider

Aging server and storage hardware had become the constraint. As the environment grew, the limits of the existing platform started showing up as slow logins, provisioning delays, and a maintenance burden that got heavier every quarter.

The complication was that clinical staff depended on these systems every working hour. In healthcare, an outage isn't an inconvenience to be scheduled around — it interrupts patient care. Any modernization had to happen underneath a running operation.

What we found

The assessment identified a refresh that had to accomplish several things at once:

• Replace server and storage hardware that was past its useful life
• Consolidate and modernize the virtualization platform
• Fix login and application performance for virtual desktop users
• Move production workloads without disrupting clinical work
• Leave behind a platform that could absorb growth rather than being outgrown again

That last point mattered. Replacing hardware with equivalent hardware would have solved the immediate problem and recreated it in four years.

What we built

We designed and deployed a Cisco HyperFlex hyperconverged platform, combining compute, storage, networking, and virtualization into one system. Hyperconverged was the right call here specifically because it collapses the management overhead — an IT team stretched across clinical support doesn't have spare capacity to operate separate compute, storage, and virtualization stacks.

The build:

• Consolidated 10 physical servers into 4 Cisco HyperFlex nodes
• Deployed a 20 TB usable all-flash storage platform
• Migrated 20 production virtual servers
• Rebuilt and deployed 200 Citrix Virtual Apps and Desktops VDI desktops
• Validated applications, storage, networking, and user connectivity before any production cutover

The validation step is what made zero downtime achievable. Every workload was proven on the new platform before anything moved, so the cutover was a scheduled transition rather than a hoped-for one.

The outcome

• Zero production downtime through the entire migration
• 10 days from hardware delivery to full production deployment
• 60% improvement in virtual desktop provisioning and user login times
• Ten servers' worth of footprint, power, and cooling reduced to four nodes
• All-flash storage replacing the performance bottleneck the users were feeling
• A highly available platform with headroom for expansion

The 60% login improvement is the number staff actually noticed. Provisioning and login time is what a clinician experiences as "the computer is slow" — and on a shift where that happens dozens of times, it compounds.

Technologies

Cisco HyperFlex (HCI) · Cisco UCS · VMware vSphere · Citrix Virtual Apps and Desktops · Enterprise all-flash storage · Windows Server · Cisco networking · High-availability infrastructure

Why it worked

Complex infrastructure migrations fail on planning far more often than on technology. Proving every workload on the new platform first, sequencing the cutover, and keeping a rollback path at each stage is what turns a high-risk weekend into a controlled transition.

The organization now runs roughly 200 virtual desktop users on a platform that's faster, smaller, simpler to manage, and built to grow into rather than out of.