Application Integration Scaling

A hub-and-spoke application integration architecture can substantially reduce the number of required integration points, simplifying engineering and maintenance. In these simplified scenarios, a business requires data collection and synchronization among a variety of disconnected, proprietary systems. For example:

• Customer-facing representatives working in a customer relationship management (CRM) system must be able to view availability of parts (in the Inventory System) to make delivery commitments to customers
• A regional sales office that tracks financial transactions in a local accounting system (labeled Regional Accounting System in the illustration) must regularly synchronize its data with the corresponding data in the company’s corporate accounting system (labeled Central Accounting System), which in turn pulls fixed-asset data from the Inventory System.

Many early enterprise application integration (EAI) architectures focused on point-to-point solutions like that shown at left in the illustration. In point-to-point solutions, developers create proprietary connection components and logic to facilitate communication between pairs of systems. In such architectures, the numbers of connection components grow faster than the number of systems being integrated. For example, three systems require three connection components; adding one more system (for a total of four) requires adding three more connection components (a total of six). Although generally not a problem for small numbers of systems, this can quickly become an overwhelming engineering and code-maintenance problem as the number of systems being integrated increases: six systems could require as many as 15 connection components, for instance.