Font Size: a A A

Designing And Implementation Of Data Integration Middleware

Posted on:2010-09-09Degree:MasterType:Thesis
Country:ChinaCandidate:P ChenFull Text:PDF
GTID:2178360278966351Subject:Software engineering
Abstract/Summary:PDF Full Text Request
We propose a middleware-based NSDI for the information system of National Audit Office to support interoperation and data-sharing between heterogeneous data sources. Information systems of different audited institutions have their own data formats and access processes. Heterogeneity of local information systems becomes a severe problem while integrated into a national wide auditing system. Issues related to maintenance and upgrades also bring challenges in building national wide auditing system.To effectively integrate heterogeneous information systems from audited institutes, we proposed a middleware-based infrastructure, NSDI, to setup a platform for a national wide audit information system. Our three-tiered middleware structure consists of presentation (frontend module), business logic (middle module), and data processing (backend module). Data processing module collects and stores data from audited institutes. Business logic module transforms data into a standard format of national auditing systems. Presentation module filters needed data and distributes data to auditors.The middleware-based infrastructure is based on our analysis of principles in data integration. Compared with traditional two-tiered client/server architecture, three tiered architecture is more effective in supporting heterogeneous data resources, separating data presentation and data storage format, and balancing workload. Object-oriented design can also make maintenance and upgrading much easier.The design and implementation of our middleware-based NDSI system is also described in this paper. Our system consists of standard data format definition module, format conversion module, file management module, data type analysis module, data structure verification module, data cleaning module, text query module, data extraction module, data loading module, task scheduling module, and activity logging module. Since these modules are designed with modularity principle, they can be re-used to construct other similar applications.
Keywords/Search Tags:data integration, middleware, NSDI, heterogeneous data resources, client/server
PDF Full Text Request
Related items