TradeWindow
A monolith connecting everything needed to move international shipments out of New Zealand, and a question: can this be modernised without a disruptive full rewrite? I analysed it, designed a phased approach — and reported that the job was substantially larger than anyone had estimated.
Context
TradeWindow is a New Zealand digital-trade software company serving exporters, importers, freight forwarders and customs brokers, connecting trade documentation and logistics processes with supply-chain participants. The system I assessed was the monolith at the centre of that.
What I did
- Analysed the monolith and identified its core system domains and components.
- Mapped the relevant data flows and integration boundaries.
- Developed the approach collaboratively with the internal development team and several TradeWindow stakeholders, rather than delivering a verdict from outside.
- Framed the scalability and maintainability trade-offs of the transition in terms the business could weigh.
The approach
A phased extraction, deliberately unglamorous and deliberately reversible:
- Move one responsibility into a new service.
- Run it in parallel with the monolith.
- Compare outputs until they agree.
- Transfer production responsibility only after validation.
- Then repeat, with what you learned.
Outcome — stated plainly
The initial analysis showed the required modernisation was substantially larger than originally estimated. My deliverable was a high-level plan for approaching the migration safely, not a service-by-service implementation roadmap. The project stopped before the team selected the first extraction candidate, so implementation never proceeded and no migration was delivered.
I include this case study precisely because of that. An assessment whose honest answer is "this is bigger than you think, and here is how you would find out cheaply" is worth more to a client than a confident plan that fails in month four.