| With the construction of the e-government application systems, the types of applicationsthat have been built in various government organs and institutions, have generated a hugeamount of business data. While in the building and running processes of the systems, becauseof the differences of the construction time, or the expanding business development, numbersof systems which have been built or under construction will cause problems. During theseprocesses, for the lack of a reasonable overall planning, resulting in the information exchangebetween the various systems is not smooth, the original business system will be unable to linkwith each other by the different platforms or architectures. Data cannot be fully shared andtake advantages of each type of businesses. And so there will be a lot of silos of information.It severely limit the further development of the e-government system. For this situation, theindustry forms for data integration in a variety of research topics, and brings out a lot ofsolutions and technologies for this topic. The SOA(Service-Oriented Architecture) andmiddleware technology will be good ideas to solve the problem. And around these technicalarchitectures, many specific technologies such as Web Service, XML (eXtensible MarkupLanguage) will also be continuous innovative. After that, the excellent design patterns, or thesystem layered architecture design philosophy, can provide support and help to improve theflexibility and maintainability of the data integration system.The research in this article is based on an actual heterogeneous data integration project,on several data integration technology, designed and implemented in accordance with theheterogeneous status between the systems, developing a complete set of data integration system, to achieve the basic functional requirements of the project. And the innovation pointof the research topic of this article is to be able to fit through the cross combination ofdifferent technologies to form a complete set of the actual system environment, dataintegration architecture and solutions, aimed at solving the highlighted several difficultintegration problems during the project implementation process. First we form a "dual-mode"design through the use of two different synchronous modes. It will respectively play theadvantages of the two modes of data integration, and make up for the lack of both sides, tosolve the system requirements while meeting the reliability and real-time issues. Second,through the integration and utilization with several design patterns, such as observer patternor adapter pattern, we design a real-time data synchronization engine based on the eventpublishing-subscription model. It uses the special message transmission process between thedata source and observer system, to effectively solve the management problem for thereal-time data synchronization among several downstream systems. It takes fully advantagesof SOA, and increases the robustness and maintainability of the system. Last, to deal withperformance issues in the system commissioning process, we specifically designed twodifferent algorithms for the more complex needs which is the full-synchronous process. Byusing agent analysis mechanisms dissecting the complexity of the two algorithms, to choose amore efficient one, integrating theory with practice, the difficulty of performance issues hasactually been solved in the large amount data synchronization situation.Throughout the course of the study of the subject, it provide an optimized design ideas,effective integration of the various data integrated cutting-edge technology, and create aspecial program to meet the data integration needs in the actual case, and finally run throughthe actual projects those are in service, have verified the use of a combination of technologies,and also have proved the value of this research. Analysis of the complexity of the algorithmalso provides a more advanced analysis method in the software optimization courses. That’sthe significance of this topic, too. |
PDF Full Text Request