Datacenter exit for an NHS hospital group

The client is an NHS hospital group running around 900 clinical workloads across two ageing, end-of-life datacentres. In a clinical setting, system availability is a patient-safety concern, not just an IT metric.

They faced a hard deadline to exit both sites, but could not accept any downtime on the systems clinicians rely on to deliver care.

Two hospital datacentres were end-of-life, with hardware failures becoming a question of when, not if. Staying put was itself a growing risk to continuity of care.

No clinical system could go dark without risking patient care, so a conventional big-bang migration was unthinkable. The tolerance for disruption on patient-facing systems was effectively zero.

The group needed to exit both sites on a deadline while keeping 900 interdependent clinical workloads running throughout. Speed and absolute continuity were both non-negotiable.

We grouped workloads into clinically-safe migration waves, mapping dependencies so no cutover stranded a system a clinician relied on. Understanding the clinical interdependencies first is what made zero-downtime achievable.

Each wave was rehearsed end to end with a tested rollback, so we always had a way back if a cutover misbehaved. Moves ran during agreed low-risk windows with clinical sign-off, keeping clinicians in control of timing.

The whole programme stayed aligned to DSPT throughout, and patient-facing systems were sequenced last, once the pattern was proven on lower-risk workloads. The riskiest moves happened only after the approach had been de-risked on safer ones.