SSAS Tabular Cubes
This release migrates from the previous SSAS Multi-Dimension (MD) to the new Tabular cubes. This required a structural re-design of the MD calculations and cubes into the SSAS Tabular structure allowing for more dynamical calculations (i.e. expected rates) at time of queries rather than pre-processing of results during cube processing. The new Tabular design is structured around a single cube called “Model” with perspectives built around this cube as business applications. This single cube combines all eleven of the previous MD cubes into only one. The object-based application hierarchy has been re-aligned within the Tabular cube structure. There are now twenty-one business applications streamlined from the previous sixty-one MD applications. The Agent transaction and movement applications have been de-commissioned. The standard naming convention of the Tabular cube will be “iw<VN>IdsTabCube” (i.e. iw100idsTabCube for this release).
Experience Studies
The Expected rates calculations are now coded in the Tabular database in DAX language instead of in SQL in the relationship DW database and in MD calculations. Across all Studies, both the initial and central rate exposure methods are now available at the same time, on either an annual or monthly rate basis. Previously in MD, the two methods were not simultaneously available in the same instance and required a custom release of the chosen method. All measure names have been revamped and standardized in order to ensure uniqueness of each name, a requirement of the single cube structure. Certain measures no longer needed have been de-commissioned (i.e. third central moment and skewness) including variance which is calculated directly in the standard deviation. The lowest level of the Policy Term Reason dimension (Pol Term Reason) has been removed resulting in a two-level hierarchy of Pol Term Reason Line and Pol Term Reason Group. Both actual and expected events are applied to the lower level Pol Term Reason Group. The lowest level of the Claim Closed Reason dimension (Closed Reason Dtl) has been removed resulting in a two-level hierarchy of Claim Closed Reason Group and Claim Closed Reason Name. Both actual and expected events are applied to the lower level Claim Closed Reason Name.
For the Policy Termination study, additional changes include:
- Substandard rates no longer applied by Policy/CoverageID. The ratings are to be directly applied to the expected rates using separate study groups.
- Joint life functionality is available for 2 lives only and the joint lives must now be in the same Study Group. Joint lives 3, 4, and 5 have been removed.
- Date Factors tables used for expected rate improvements must now be specified at the calendar year level (i.e. set DateMthFromKey = YYYY01 and DateMthToKey = YYYY12).
- For the Annual central rate (ACR) method, the Date Factors are now applied directly to the annual initial rates. Previously, they were applied to the monthly central rate as an approximation.
Performance and Design
SQL Server 2016 Platform
This release with prefix “v10.0” was developed on the SQL Server 2016 Enterprise Edition with SP2 and on SSAS Tabular mode. For SSAS Tabular, the server specification requirements are very different from SSAS Multi-dimensional and clients who are upgrading should re-evaluate their hardware needs.
Project Deployment Model
For SSIS projects, the SSIS packages has been migrated to the “Project deployment model” from the previous “Package deployment model”.
The release package folder now stores just one file “iwPckg.ispac” that contains all packages. This is deployed to a single Integration Service Catalog on an instance of SQL server.
The SSIS project uses Project Parameters to specify package properties including server and database names. With this change, the Insight iwMaster database is no longer needed and has been de-commissioned.