There are two modes to develop.
Tabular Models where you have tables and relationships.
Benefits: Fast by design and is in memory
Easy to start up and is simple to use.
Multi Dimensional Models:
Since SSAS 2000. It has Dimensions and Measure Groups. It is very mature and is complex for development and use. You can use both methods and there is no better way. It depends on the environment and the problem that is being tackled.
Tabular model can be connected to multiple data sources. Where as Multi Dimensional models is attached to Relational only. The Data Access is also different for the different Data models in SSAS 2016.
For Multi dimensional its MDX and for Tabular its DAX to query the models.
Themes for the SQL Server 2016.
To use through directQuery without Storing data in Tabular Models.
New Features in SSAS:
Most new features for 2016 is for Tabular model and there are no new SSAS data mining features or updates.
Analysis Services Management Objects (AMO) has been refactored to include a second assembly.
The DBCC for analysis services runs internally to detect potential data corruption issues on Database Load.
Extended events can now be monitored in real time within SSMS and can be stored for offline analysis. Computers on the network can now be made DBAs in SSMS 2016.
