Design And Development Of Data Integration Middleware For Multi-source Heterogeneous Data

With the competition from individual enterprises evolution to the supply chain and supply chain,as well as the importance of big data becoming obvious,the big data technology supporting the enterprise group precision management and decision-making ability of the industrial chain becomes an important research direction for automobile industry.The quality and effectiveness of big data analysis depend on a great extent on the integration and integration quality of the data in the industrial chain.But in the domestic automobile industry chain,the cooperation data of industry chain with automobile manufacturing enterprise as the core has incomplete data and data multi-source heterogeneous two major problems,and these two problems are precisely the cross enterprise,cross-sectoral industry chain collaborative data integration and integration of the difficult problem in the distribution,autonomous environment.The integration and sharing of multi-source and heterogeneous business data in the industrial chain becomes the key issue of big data analysis in automobile industry.Therefore,based on the research of the technology support project of Sichuan Province,"the third party manufacturing service cloud platform supporting technology for automobile service lifecycle(SLM)"(2015GZ0076)and "distributed resource giant system and resource synergy theory"(2017YFB1400301).,a distributed autonomous multi-source heterogeneous data integration solution on the basis of in-depth research on data integration technology is put forward.The key technologies are studied,and the corresponding cross-platform data integration middleware is developed to support the integration of distributed multi-source heterogeneous data in the industrial chain.The main tasks of this article include:(1)According to the existing business data interaction characteristics of automobile industry chain and the specific needs of data integration,based on the research of data integration technology at home and abroad,the idea of data integration on the basis of middleware is put forward,and analyzes the key requirements of multi-source heterogeneous data integration middleware,next,The function design,overall design and detailed design of data integration middleware are completed.(2)The two key core issues in data integration: multi-mode conflict resolution problem of multi-source heterogeneous data and consistent transparency ofcommunication and access among distributed autonomous data sources in heterogeneous mode,a pattern tree method is proposed to describe the common data pattern of heterogeneous data and to solve the heterogeneous conflict problem by tree transformation.,a unified and common method of request execution communication is proposed,the algorithm of query request verification,decomposition and dispatch is proposed,which realizes the request data source in a uniform way,and a dynamic adapter management method is proposed,optimizes the traditional architecture of "one data source adapter",reduces the cost of adapter maintenance and expands the scope of application.A unified data representation method based on cross-platform XML is proposed,and completes data migration and reorganization.(3)using the platform-independent JAVA and XML technology,the heterogeneous data integration middleware supporting the automobile industry chain collaboration platform is developed and applied in the automobile industry chain cooperation platform.As the middleware designed in this paper has the characteristics of platform independence and configuration flexibility,the work of this paper can be extended to other distributed multi-source heterogeneous environments in addition to its application in the environment of similar automobile industry chain cooperation platform.