Land every migrating workload
on a target chosen for it.

A source-to-target map records, for every source component, the target service it lands on, its disposition and the tooling lane that moves it. The dispositions are retire, carry over, convert or redesign. The map is the contract for the wave plan: the waves are sized and priced from it, and it shows which work is repeatable and which needs engineering judgment before conversion starts.

This work starts after the platform is chosen. Choosing between platforms is Technology Selection in Strategy & Advisory. Here the platform is agreed and the move is funded, and the question is where each component belongs on it: which extraction job, warehouse table, semantic model or low-latency query lands on which Fabric or Databricks service, and why that service fits its job.

A one-for-one translation carries the old limits, and the dead workloads, into the new estate. An extraction job, a warehouse table, a semantic model and a low-latency query do not belong on the same service just because they share a source system. We read each component for its dependencies, latency, access and volume, and retire anything with no continuing purpose before it costs conversion effort.

The duration is scoped in the first working session, because it depends on the size of the source inventory, and the price is fixed once it is scoped. You keep the source-to-target map, with a target service and disposition for every in-scope component, and the tooling and dependency plan that feeds the migration wave plan, whoever runs the waves.

You need an agreed platform, a funded move and a source inventory to work from. If the platform is still open, start with Technology Selection in Strategy & Advisory. The first step reads that inventory and groups the source databases, pipelines, report models and serving workloads by dependency and operating need, so the map starts from how the estate actually fits together.