Forty workloads out of one data centre, before the lease ended.

The brief was non-negotiable: lease ends June 30, no extension. The team had been planning the migration for eight months and not migrating anything because every plan got bigger when reviewed. We started by drawing a line through workloads — rehost where we could, replatform where it made sense, retire where nobody could explain the workload’s purpose.

AWS Application Migration Service did the bulk of the rehost work. Twelve workloads got the lift-and-shift treatment first to de-risk the schedule. DMS handled the database tier with continuous replication and tight cutover windows. Direct Connect bridged the colo to the AWS landing zone during the transition.

The last workload cut over thirteen days before the lease ended. The colo team did the physical decommissioning under contract. Total downtime across all 40 workloads, aggregated: under four hours. Run-rate cost on AWS came in 14% under the colo total cost of ownership.