Based On Soa Architecture Design Of Heterogeneous Data Source Wrapper

With the rapid increase of the information in production and our daily life makes the need of data not only limited to local data resource, but also the other data. The isomerous they have makes corporation, department and organization inconvenience to call on for each other, and reduce the percent of information utilization.This dissertation is supported by The National Development and Innovation committee High-Tech Industrialization Project: The SynchroESB Software System which is a part of project Synchr Business Integrator Platform. This system realized the operation of application soft system and isomerous data resource one another, makes users conveniently, fast and exactly get data from isomerous data resource and have no use for understanding material information structure and accessing interface of exact data resource.The wrapper which is used to encapsulation for isomerous data resource is an important part of it. It makes the outer application system link to service bus, and can interchange of other program section and application system with standard message. It provide wide connect to database, message server and carry-over system, so we in-depth study on it in this paper.Firstly, we discuss the technology corresponding on XML, the basic theory of structured and unstructured isomerous data resource and the concept of Service-Oriented Architecture and Enterprise Service Bus in this paper.Secondly, we described the system frame of SynchroESB software platform which is service oriented. Based on this, we analyze the exact deployment and implement by a concrete applied scene in detail. Then we discussed the aim, effect and existing form of wrapper in SynchroESB system.Then, we described the design and key technique of the database of structured data source. The techniques include XML file analysis, mapping between related database and XML and change from XPath to XML. Based on analyzing the characteristic of web page and three techniques of web information extraction, we present a new design blue print of web wrapper for unstructured data source. It realized semi-roboticized web information extraction by web page define, Objective model definition and model learning.At last, we make a further discuss and give out a prospect of wrapper of data source.