Supply Chain Technology

Although some organizations are using new application integration-enabled or traditional middleware to tie custom systems together with those of their supply chain members, many others are looking to ERP packages as points of integration. ERP provides members in the supply chain with a common application layer to exchange information through well-defined interfaces, business logic, and repositories. Unfortunately, getting all members of the chain to run the same ERP application or any other application, for that matter is an almost impossible task.

ERP systems are designed to provide their owners with a system for planning, controlling, and monitoring an organization's business process. What's more, ERP systems offer easy, system-to-system integration through standard mechanisms for communication. They also build a common understanding of what the data represents and a common set of rules (methods) for accessing the data. However, traditional ERP systems are not necessarily supply chain systems. Supply chain systems must deal with the complexity of integrating information from any number of disparate systems that span multiple organizations. Although some properties of ERP systems are appropriate for use with the supply chain, others are not. Typically, ERP systems exist within a single organization, using a single database as the point of integration with internal organizations. As a result, there is a single database and network protocol standard; thus the enterprise system is able to work with other departmental systems without having to deal with complex integration problems.

When supply chains integrate external systems, it is nearly impossible to get all member organizations to agree on a set of standards (e.g., network protocol, database architecture, and process integration). Because of this difficulty in arriving at a unanimous agreement, the best solution is moving information from system to system using layers and layers of middleware, gateways, and adapter technology. In most cases, very little processing can be shared among the members. As a result, the supply chain system architecture can grow exceedingly complex and confusing.

Because of this reality, two problems need to be solved. The first is getting all member organizations to agree upon a single communication standard and a single set of application semantics. Although it might appear that communication would be an easy problem to solve, it rarely is. Some of the member organizations, especially the smaller manufacturing concerns, face significant costs to get their systems ready to link to others in the chain. Once everyone is ready, the second problem must be addressed all members need to agree upon a common communication mechanism and middleware approach, allowing all member systems to seamlessly access processes and data in the extended enterprise.

As difficult as these two problems might appear, implementing the solutions is often more difficult.

These difficulties will not be overcome anytime soon. The value that the new application integration technology brings to this problem domain is the ability to integrate many of these systems without requiring that significant changes be made in the systems owned by any of the trading partners. Instead of an organization having to change the database, the application, and possibly even the network, application integration-enabled middleware can adapt each system into the larger system by accounting for the differences in data and methods within a middle tier (see Figure 17.5).

Figure 17.5. Application integration technology is able to link many different systems in the supply chain by not requiring that changes be made to the source or target systems.

The middleware best suited for supply chain systems is represented by integration servers and application servers. Because message brokers are message oriented and asynchronous, they're able to move information from system to system without requiring that all systems be up and running at the same time.

Application semantics is another issue that organizations must confront. Although most organizations can agree upon relational databases, few can agree upon a common schema and architecture that can be shared across the chain. Supply chains solve this problem either by forcing a common data architecture among the members, or by using message brokers or data translation middle ware to adapt the data to various models and schemas. The latter is, as in other contexts, the path of least resistance.