A Unified Method Of Heterogeneous Data Resources In Network Environment

The development of Internet makes enterprise can utilize more volumes and types of information than ever before. But it also brings new challenges to information access. More and more semi-structured or unstructured information is used in enterprise applications such as relational database, XML, Web Service, format text and so on. It makes access and management of heterogeneous data more complex. Due to complexity, diversity and distribution of Internet-based resources, there is not a satisfactory solution yet. This paper relies on "XBus data aggregation platform" and researches on some difficult issues, which includes access of heterogeneous data sources, encapsulation of data sources and integration dynamically of data sources in data layer. The main work and contributions are as follows:1. A unified method of resources.First of all, the range and classifies of data resources are defined. Then the data source is defined formally. Secondly, a basic data structure based on nested relational model is proposed, which unifies variable information representation forms into nested relation. Because of the uniform abstraction of information, the complexity of accessing variable resources is reduced. Finally, some rules of model constructing for different types of data sources are given.2. An encapsulation method of data source.First of all, a method based on user description is proposed which can transform data source into uniform model. A group of wrappers are used to extract data from different data source. The data source encapsulation is divided into two phases:getting metadata information of data source by interacting with the users, then generating the initial schema of data source; comparing the predefined schema to all schemas in model library, then according to the result to select that reusing existing wrapper or re-creation a new one. Secondly, an algorithm of model matching in model library and principles followed by model selection are given.3. These methods have been designed and realized.A system of heterogeneous resource integration has been realized, which is a subsystem lied in data layer. Thus, experiments are designed based on supporting environment "XBus data aggregation platform". These experiments illustrate the effectiveness and the usability of the system.