To simplify the time-consuming Microsoft Excel-based process, our engineers introduced Supply Protection (SUP) flagging, a system that tracks alterations in product attributes via a flagging mechanism and relays this information to downstream systems.
Supply protection is a feature of the Advanced Available-to-Promise business function in SAP S/4HANA. The transition to SUP flagging diminishes the likelihood of data inaccuracies and eliminates the technical challenges associated with manual updates.
The company’s inventory planning group will use the SUP Flag information to filter eligible products from Demand Consolidation and Disaggregation (DCD) to Protect, Chase, and Cancel outputs, creating and maintaining SUP buckets in SAP/S4HANA.
A third-party partner supplies SUP flagging-related data as files stored in AWS S3 buckets. Based on the product information provided, data is then transferred to DCD. This becomes a source of truth for further downstream systems like inventory planning and product dimensional data systems to safeguard SUP-flagged products.
Our engineers created Directed Acyclic Graphs (DAGs), a data structure that models dependencies in complex planning processes. The DAGs allow for efficient computation and analysis by ensuring tasks are executed in a specific order without circular dependencies. They enable the system to self-trigger whenever a new file or data is uploaded.
Here are a few key takeaways from the solution:
- Automatic tracking reduces time and errors: Using the SAP Support Protection flagging allows our solution to automatically track alterations in product attributes, saving time and reducing errors.
- Single source of truth for SUP flagging: In the new system architecture, DCD is the single source for SUP flagging across all teams.
- Separation of tasks improves efficiency: Since the DAGs for triggering the code and making the changes are provided to the team, they can be self-driven by the team with no dependency on the development team.
- Test data prototypes save time: Since getting data from the source is usually tedious, prototypes with test data for all possible EDGE cases were written and validated in advance.