...
- Logical data separation, no confusion, no necessity for explicit description why configuration and data are stored at same table
- Performance improvement, due to number of anchor records (expected) is significantly lower then the amount of data fragments
- No extra complexity in data processing (CRUD implementation) and dependency management (FK constraints)
Cons: none
- Some refactoring work
Mitigation: make decision soon and act on up to prevent it becoming more difficult
...