Single large semantic model versus multiple semantic model, which is better?

our student data warehouse consists of multiple star schema models (with conformed dimensions) serve of basic constituencies, example:

• admissions & academics
• financial aid
• student accounts
• career services
• etc

we building tabular bi semantic model/s these modules, , in process, i'm trying analyse if better create single large semantic model constitutes of above said modules put together, or create multiple semantic models (separate ssas tabular databases) caters individual modules considering activities, business processes, incremental load schedule different. thing these modules have in common of conformed dimensions such student, program, school campus, etc.

what best practices, key considerations, , concerns if 1 has come across such scenarios? please advice, appreciated. 

this depends on end user requires data. since mentioned conformed dimensions, means slicers , dicers can access measures different star schema fact tables.

if users need measure sales 1 schema next inventory flow in schema product, need 1 model.

what happens in experience multiple star schemas in 1 model not dimensions conformed. so, user might place dimension attribute next measure not related , think model not work because result confusing. requires educate end users can sliced , can not.

it not wrong make multiple models, have think requirements need satisfied end users.

---

thomas leblanc twitter ( @thesmilingdba )

SQL Server  >  SQL Server Analysis Services