Legacy ERP architecture offers an incredibly, efficient means of operational resource management, but a real challenge comes from extracting business insights from them. Over the past 30 years, ERP data system, such as SAP, can be hard to interact with especially at the source database level. Whether initial translation of business logic and hierarchies create significant customizations, as well as, merging those changes into analytical applications. Overall, the entire process of designing self-service reporting with business level context can be quite cumbersome, looking at an example platform like, SAP, which contain pre-packaged modules (MM, SD, PP, etc), integrating these systems into a series of pre-built analytics.
The orchestration and integration over a wide range of open source technology solutions with some commercial CDC and reporting solutions into a reference solution that mimics several real customer scenarios today, living on relational platforms. Key considerations of extracting from the operational system of record, especially the merging of multiple systems in different time zones, will be addressed. Furthermore, the integration concerns of an analytics Hadoop platform, using HIVE Acid and Merge, as well as, flattening techniques for dimensional models. Many times a customer is temporarily limited in the range of data their ERP can contain, and older data is often offloaded to secondary systems or cold archiving entities. That goes away, but the opportunities expanded with real-time reporting across all of history, and the expanded use cases with advanced machine learnings methods.