Previously, users and application developers, through manuals and experience, understood the context of the few systems that served their portion of the enterprise. But the number, types, and increased scope of data and systems that are now being integrated make it impossible to expect users to understand and remain current on the context or meaning of all information. Our research has the goal of representing, moving, and processing the context along with the information it describes. This requires both representations, models, manipulation languages and reconciliation algorithms for context knowledge. In this position paper we present a number of significant research areas which we believe must be resolved so that context knowledge can be used to simplify the integration of multiple disparate database systems. As shown in Figure 1, the export context defines the meaning of the data provided by a data source while the import context defines the context requirements for the data receiver. Providing this context knowledge requires an understanding of issues in context representation, context models, common metadata vocabularies, comparisons of contexts including transformations, and system's architectures and operations.