Modernisation path
A path you can audit
This page outlines how we typically sequence legacy application modernisation work. Your programme may reorder steps when regulatory or seasonal freeze windows demand it.
1 — Establish what production actually does
We interview operators and read scheduler logs, not only product backlogs. The output is a runtime inventory: languages, binaries, batch windows, and hidden scripts that touch the same tables.
2 — Draw strangler seams
We mark boundaries where read models, event topics, or façade APIs can absorb traffic without rewriting every screen. Each seam includes a parity test plan before code changes land.
3 — Prove data parity
Sampled transactions are compared between legacy and candidate services. Checksum reports are versioned so internal audit or risk teams can replay the evidence.
4 — Cut over with rollback
Maintenance windows get explicit rollback triggers — not a slide that says “revert if needed.” We stay available through the window when the statement of work includes cutover support.
5 — Stabilise and hand over
Runbooks, on-call playbooks, and knowledge-transfer sessions are part of the final milestone. Advisory retainers can continue for quarterly architecture reviews.