Cloud migration and legacy modernization
From old servers and monoliths to the cloud or Kubernetes, while the business keeps running.
For teams with systems nobody wants to touch: on-prem databases, monoliths, hand-built servers, or two tools that must stay in sync. I analyse the code with AI help, plan the move in small steps, run old and new side by side and cut over only when row counts and checksums match.
Sounds familiar?
- A critical system runs on hardware or software that is out of support
- Nobody fully understands the old code, and the people who wrote it are gone
- A big-bang cut-over is too risky, but standing still costs more every year
- Regulated workloads cannot loosen a single control during the move
How I work on it
- 1
Understand what is there
AI-assisted code analysis maps dependencies, data flows and hidden scheduled jobs. Every finding is checked against the running system before it goes into the plan.
- 2
A plan in small steps
Each workload gets its target, cloud, Kubernetes or staying where it is, and each step has a way back. No step depends on everything else being finished.
- 3
Landing zone as code
The target platform is built in Terraform or Bicep with strict Dev, Test and Prod separation and security baselines enforced as code, so the new home is safer than the old one from day one.
- 4
Run old and new in parallel
Data is synchronised while both systems stay live, as in a two-way ERP and CRM sync where nobody had to stop working. Differences show up in reports, not in customer complaints.
- 5
Cut over on proof
The switch happens in a quiet window, after row counts and checksums match, with every client of the old system inventoried and a tested way back. Then the old system is switched off for good.
What you get
- An inventory of systems, dependencies and data flows, checked against production
- A migration plan in steps, each with risk, timing and the way back
- A landing zone in Terraform or Bicep with security baselines as code
- Data sync and parallel run, with row-count and checksum reports
- Cut-over runbook, decommissioning of the old system and a handover to your team
Real projects
- Automotive manufacturer, 2022–2023A regulated workload moves to Azure, fully in Terraform~25%
- Professional services, ERP and CRM, 2019–2022Two-way ERP ⇄ CRM sync while both systems stayed live2-way
- Public sector, healthcare, 2019–2022Sensitive health records between two public systems, none lost0
- Media-technology platform, 2026Hybrid Kubernetes for live broadcast130
Questions buyers ask
Do we have to move everything to the cloud?
No. Each workload goes where it fits best: public cloud, your own data center or both, run as one platform. Some systems are better left where they are and only modernized.
How do you avoid downtime?
By running old and new in parallel with synchronised data, cutting over in a quiet window only when counts and checksums match, and keeping a tested way back until the new system has proven itself.
How does AI help with old code?
It reads large codebases fast and maps dependencies, data access and business rules. I check every finding against the running system, so the plan rests on facts, not on what the AI assumed.
What about compliance during the move?
Controls move with the workload, as code. In a regulated automotive migration not a single control was loosened, and security baselines were enforced for every team.
How do you price it?
After a 30-minute call: a fixed scope for the assessment and plan, then a fixed scope or a few days per month for the migration itself.
Read more
- From AI prototype to production: the eight layers I add, in order
- Air-gapped systems: running critical software with no internet
- Right-size Kubernetes: the playbook that took a cluster from 7 to 4 nodes
- The legacy migration playbook: waves, parallel runs, and identical row counts and checksums
- A landing zone for mid-sized companies, in Terraform or Bicep: the parts that matter
- Restore tests: a backup you have never restored is only a hope
Get in touch
Tell me what you're building. I'll reply within 1–2 business days.
- Company
- MADO Consulting
- Follow me on LinkedIn
Available worldwide, remote. On-site on request, depending on the country.
Opens Calendly's booking calendar here. Loading it sends data to Calendly (see Impressum).