| Today, with the high-speed development of information technology and network technology, it's more and more necessary for enterprises to optimize there external and internal business flow management and realize the automation of the data flow and business operations. To be adapted to the enterprises' requirement, more and more application system are developed and used and there is more and more available information around us while more and more people realize these datum's value, but these resources cannot be used availably—these datum spread over many heterogeneousdatabases and many heterogeneous systems with different formats. It's very hard to collect them to analyze and exhibit. In these conditions, cross-enterprise data integration emerges as the times require.The first part of this paper points out the meaning of data integration, analyzes some existing data integration solutions and sum up problems to be solved in a nicer data integration solution. The next part introduces some theories and technologies as the foundation of data integration. Then a basic data integration model is presented. It is a C/S model composed of an exchange server and several adapters. The exchange server is in charge of managing all resource metadata information. Adapters as the integration platform agent for local applications exchange are in charge of finishing the data formats mapping between the local data source and virtual database represented by the platform and exchanging resource metadata information and resource information with the exchange server. Later, some important problem's solutions of the basic model are provided. The following extends and perfects the basic data integration model from several different points of view such as the link establishment, the multiserver model, the global datum and views handling and the semantic model. The paper's last part gives three typical examples based on this model to validate its feasibility and practicability.
|
PDF Full Text Request