✓What You'll Learn
A manufacturing company with 180 applications across two ageing data centres needed to move to cloud. 90 days later, 60 applications were running on AWS, infrastructure costs were projected down 34%.
When a £250M revenue manufacturing company came to Diztaly, they were running 180 business applications across two ageing data centres — with infrastructure refresh costs of £8M on the horizon and a board mandate to "do something about IT." Ninety days later, their first 60 applications were running on AWS, their infrastructure costs were projected to fall by 34% over three years, and their IT team had been freed from data centre management to focus on digital transformation initiatives that were generating measurable revenue impact. This is how we did it.
The Starting Point
The client's IT estate was heterogeneous: a mix of Windows Server and Linux workloads, a legacy ERP on-premise, modern SaaS applications, and several custom-built operational technology systems. Some applications had not been touched in eight years. Others were business-critical systems that required near-zero downtime during migration. The board mandate was clear — reduce infrastructure costs and improve resilience — but the timeline was ambitious: first migrations within 90 days.
The 90-Day Migration Programme
We began with a two-week application discovery and dependency mapping exercise, using automated tooling to catalogue all 180 applications, map their interdependencies, and assess their cloud readiness against the 5R framework (Rehost, Replatform, Refactor, Repurchase, Retire). The assessment revealed: 68 applications suitable for direct rehosting, 42 applications requiring minor replatforming, 24 applications to be replaced with SaaS equivalents, 31 applications to be retired, and 15 requiring refactoring — deferred to a later phase.
Migration Execution
We structured the migration in three waves. Wave 1 (Days 1–45): 40 low-risk, rehost-suitable development and test environments — building team capability and validating the migration tooling with applications where downtime risk was acceptable. Wave 2 (Days 46–75): 20 business applications of medium criticality, including several that required database schema adjustments for cloud-optimised storage. Wave 3 (Days 76–90): Phase 1 go-live, covering all Wave 1 and 2 applications in production, with legacy data centre retained in warm-standby for four weeks before decommissioning.
Results at 6 Months
| Metric | Pre-Migration | 6 Months Post | Change |
|---|---|---|---|
| Infrastructure monthly cost | £142,000 | £94,000 | -34% |
| Applications on cloud | 0 / 180 | 109 / 180 | Phase 1 complete |
| Average deployment time (new environments) | 3–4 weeks | 2 hours | -97% |
| Unplanned downtime (quarterly) | 14 hours | 1.2 hours | -91% |
| IT team time on infrastructure management | 60% of hours | 18% of hours | -70% |
Following successful migration, the team is now implementing FinOps practices to further optimise the cloud cost base and exploring cloud-native AI capabilities that were inaccessible with on-premise infrastructure.