From 12 physical servers to 3: a Ghanaian retail bank cuts its IT bill by two-thirds
The problem
Twelve aging physical servers running in a branch-headquarter closet — monthly electricity and cooling cost of GH₵ 18,000, no disaster recovery, 6-hour restore time, and two hardware failures in 18 months.
What we did
We consolidated the workloads onto three modern hosts running VMware vSphere, configured HA clusters, introduced snapshot-based backups to a remote site in Kumasi, and wrote a live-migration runbook the bank's own admins could execute.
“Flash Code's team embedded with ours for six weeks. They didn't just virtualize the servers — they taught our admins how to own the platform after they left.”
— Head of Infrastructure, client bank
How we delivered it
Step 1Discovery
Full workload inventory and dependency map over two weeks on-site.
Step 2Design
Right-sized three hosts with N+1 capacity and dual 10GbE for live migration.
Step 3Pilot
Migrated the non-critical workloads first to prove the runbook before banking hours.
Step 4Cutover
Weekend migration window with rollback to original hardware if anything failed.
Step 5Handover
Two-week shadow period and a runbook every admin could read without training.