The hardest work of all in getting an enterprise resource planning (ERP) system up and running is completing integrations to databases scattered throughout divisions, departments, and workgroups. ERP vendors may suggest that integration is painless as long as their specific database is used, but that's a fallacy; all databases have both strengths and weaknesses when integrated into ERP systems, and their relative levels of performance vary as part of a functioning and complete ERP system. Add in the fact that many companies rely on a single ERP instance across a global set of factories, distribution centers, and service centers—even creating a single system of record that coordinates transactions across joint ventures—and the complexity of coordinating databases within an ERP system becomes clear.
The goal of this article is to define the strengths and weaknesses of three types of databases—Oracle, Microsoft SQL Server, and open source databases such as PostgreSQL and MySQL AB—as they relate to globally focused ERP systems with a primary goal of consolidating order management functions. To begin, let's examine a model for distributed order management. We'll complete the analysis by assessing each database on the major system attributes in the model.
ERP Is Industrial-Strength Order Management
Business models, approaches to producing and selling products and serving customers, and techniques for managing mergers and acquisitions vary widely among manufacturers. Despite all these variations, however, all ERP systems share 14 critical system attributes. Taken together, these 14 components form the Distributed Order Management model (see Figure 1) originally defined by AMR Research in Rod Johnson's report "Consolidated Order Management—ERP Alone Doesn't Deliver" (February 28, 2003):
- Role-Based and Contextual Personalization
- Order Promising
- Sourcing Allocation
- Order Visibility
- Consolidated Settlement
- Inventory Visibility
- Automated Replenishment
- Delivery Management
- Service Management
- Event and State Management
- Order Broker (Integration Framework)
- Business Process Management
- Master Data Services
- Analytical Data Services
Figure 1 Distributed order management is the heart of an ERP system—and the core to evaluating the pros and cons of databases.
I've used this model to define the weaknesses of integration points on several enterprise application integration (EAI) applications, including Siebel's UAN and BEA Systems' applications. The problem is the ability to deliver on the promise of integrating multiple and highly dissimilar databases—in addition to troubleshooting ERP systems' database integrations with Microsoft, Oracle, open source, and homegrown systems.
Let's consider the various model components and database dependency definitions.