Our actions and decisions are only as good as the data we use to rule them. That being true, when designing complicated systems, ensure the integrity of the data set involved becomes enormously important. Obviously, the data that resides in a system should be accurate. However, it also needs to have integrity, which means consistent and not unnecessarily redundant.
A good data set is invaluable. Done well, it puts all of the data elements in one place, making it easy to share, analyze, extrapolate, and use. Which system owns the data is potentially even more important. If you want to fundamentally alter the data based on business rule or a business logic, then it should be done in the system that owns the data—so it’s critical to track the lifecycle of the system and to refer users to the primary system for specific data.
For example, sales order (SO) information is always owned by an enterprise resource planning (ERP) system, since that is where all the aspects of the order are handled and tracked throughout the lifecycle of the sale. That same sales order might be downloaded to a warehouse management system (WMS) for fulfillment. The WMS will use the order information to fulfill the order, to optimize the fulfillment process, for packaging instructions, and more. However, this system should only allow information lookup—while changes to the data should be done in the ERP.
Let me offer a real-world example. When I was working at multinational computer company, we sold computers all over the world. We had a complicated challenge of streamlining the system integrations pertaining to shipment data from various manufacturers worldwide. These suppliers would share advance shipment notifications (ASNs) through electronic data interchange (EDI). About 856 transactions in all, these provided all shipment information that allowed the recipient to be ready to receive the product. Our goal was to create a single consistent architecture so that the Global Supply Chain Systems Group could support it.
Click here to continue reading this article.
Comments
Post a Comment