Earlier this year the Oracle Fusion team introduced the new "Fusion Data Extraction" feature, which is being positioned as a replacement/improvement to BICC.
Data Extraction Using Read-Only Replicated Data
Advertised as a part of this feature is a "near real-time replication of Fusion Cloud data into a dedicated, read-only data store within Oracle Autonomous Data Warehouse". Although it doesn't appear that customers can directly access the ADW that supports this feature, it does beg the question if the FDI product can/should integrate to this ADW to supplement FDI with practical real-time data, reporting, and analytics use cases.
The argument for a denormalized schema shouldn't go a way, given that the denoramlized model is both user friendly and technically designed for faster analytics. However, for scenarios where customers demand real-time reporting and analytics, utilizing the RODS ADW layer in FDI could be a massive feature addition, especially for many customers who in general prefer to the FDI user experience, but at least today can't use it exclusively because of a lack of real-time analytics.
Questions that would need to be addressed if implemented:
(1) Would the semantic layer represent these "RODS" based data sources as completely separate subject areas?
(2) How would dashboard asset design change be impacted when using RODS based data sources? Assumption is that RODS ADW schema is more normalized, and hence less efficient for "wider" and more "cross-fuctional" analytics.
(3) In general an approach to balancing user adoption of utilizing a RODS based analysis, subject area, dashboard etc vs. the current state model.
(4) In general, should the existing data augmentation feature for Fusion sources be migrated off of BICC towards the new data extraction view object model? Given there's not 100% coverage on the transactional Fusion data yet assuming this will take time to implement.
