Integrating complex enterprise systems is challenging and reliability and performance of the integrated systems can be problematic when using typical solutions of distributed transactions or on-demand message-based querying. We describe a data-oriented approach for enterprise system integration that uses information brokering. A broker application isolates the interactions between a local enterprise application server and a wide variety of remote systems, performing data acquisition, storage, transformation, update and status management. We describe a prototype book brokering system developed using Java 2 Enterprise Edition, CORBA, Java Messaging Service and Web Services and using our integration strategy. We outline the architecture, design and implementation of this prototype and summarise our experiences with this information integration technique.

Pervasive business requires information brokers that support customer/supplier enterprise system interactions in sensible ways. We present a summary of our model for pervasive business: a virtual database used to aggregate information from parts of multiple enterprise systems. Aggregated data is typically managed by a brokering enterprise in a high performance main-memory database. Replicated information from multiple customer and supplier enterprise systems is managed by the broker. This data is acquired and updated seamlessly by the broker's Enterprise Systems Logic™, using a variety of technologies to acquire the data, translate it into a canonical form and apply updates to the originating system when appropriate. We present the motivation for this novel pervasive business architecture, a review of related systems and technologies and our work to date on this project.