Rebuilding cross-border settlement on an event-driven core
An illustrative scenario showing how a cross-border payments provider could replace a struggling monolith with an event-driven settlement core.
Illustrative solution scenario, not a client engagement.
Faster, more predictable settlement times
Headroom for higher peak throughput
An auditable, immutable settlement record
The challenge
What was breaking.
Settlement runs as a nightly batch inside a decade-old monolith. Any failure mid-run means a manual replay the following morning, and reconciliation queries lock the same database the payment API depends on.
Regulatory reporting needs an immutable audit trail the existing schema cannot produce, and a planned expansion into new corridors would multiply peak volume.
Our approach
How we delivered it.
Strangler-fig migration
Route traffic through a façade and move one payment flow at a time onto the new core, avoiding a big-bang cutover.
Event-sourced ledger
Model every balance change as an append-only event, giving the compliance team a replayable, immutable audit trail by design.
Isolated read models
Move reporting and reconciliation to projections, removing contention with the transactional path.
Load-tested release path
Run automated soak tests against forecast peak on every merge, with canary releases behind feature flags.
Services applied
The team on this work.
- Custom Software Development
- Cloud & DevOps
- System Integration
Technology
What it runs on.
- TypeScript
- Kafka
- PostgreSQL
- Kubernetes
- Terraform
- AWS
Explore more
Related work.
Tell us what needs to ship.
A 30-minute call with the engineers who would run the work — not an account manager.