The Service Layer, direct database access and the extraction lanes
Inventory every Service Layer, DI API, B1if, file and direct-database consumer, then map each onto a governed target pattern — because D365 exposes no transactional database to integrate against.
What you will be able to do
Inventory every integration and reporting consumer, including the ones nobody lists
Sequence extraction through the four source lanes in the right order
Map each source integration surface onto a governed target pattern
Replace direct database reporting with a target analytics design
Build a validation set that proves extraction rather than merely completing it
Introduction
Integration on this path has an asymmetry that shapes the whole workstream: the source system is easy to integrate with, and the target is not — deliberately.
The version facts that change the design
Three version-sensitive facts materially change what an integration design can assume.
Building the consumer inventory
The inventory is the deliverable that sizes the workstream. Build it from multiple angles, because no single source is complete.
The four extraction lanes
Extraction is a separate exercise from integration, and the playbook sequences it into four lanes. Order matters.
Mapping source surfaces to target patterns
Each source surface has a governed target counterpart, and none of them is a repoint.
Replacing direct database reporting
This is usually the largest single population in the inventory and the one users feel most.
Validating extraction
The source validation set is specific, and it is the right list because it is what finance and operations recognise as proof:
Knowledge check
Summary
Integration is where the source estate's accessibility becomes the migration's largest hidden scope.