ePrints.FRI - University of Ljubljana, Faculty of Computer and Information Science

Enhancement of BI4Dynamics Analytical System into a Business Intelligence Modular Framework

Marko Ilić (2010) Enhancement of BI4Dynamics Analytical System into a Business Intelligence Modular Framework. EngD thesis.

Download (1894Kb)


    BI4Dynamics is a business intelligence solution for the Microsoft Dynamics NAV transactional system based on the SQL Server platform and .NET Framework. Step-by-step analytical system's deployment wizard deploys a fully functional data warehouse and analytical database on the SQL Server. The NAV data is pumped into the staging area of the data warehouse by a group of ETL processes where it gets cleaned, restructured and in the end stored into the presentation area of the data warehouse. Analytical database consists of a group of dimensions and cubes, which reads the data from the data warehouse presentation area. The BI4Dynamics architecture was not designed in a way to provide a modular structured data model. Therefore, every BI4Dynamics implementation comes with six standard analytical areas, which are deployed by the deployment wizard. All additional analytical areas need to be integrated into the BI4Dynamics analytical system manually. When BI4Dynamics entered the worldwide market, the need of a modular data model, which allows deployment of any group of BI4Dynamics modules using the deployment wizard, grew stronger and stronger. In the analysis of the old BI4Dynamics data model, I described the entire architecture where I pointed out the key architectural qualities that need to be kept in the new architecture. I also pointed out some obsolete architectural parts that need to be discarded in the new architecture. The biggest changes were implemented in the data warehouse staging area architecture. The process of building the staging area in the old architecture was too complex to provide modular deployment of the data warehouse. In the new architecture, I minimized the number of setup tables and procedures which are used during the deployment of the staging area. The presentation area of the data warehouse did not undergo many architectural changes due to the nature of the dimensional data model, which provides modularity by default. The presentation area objects underwent some minor changes in the form of changing the naming convention, which is more user-friendly in terms of locating the objects on the SQL Server in a timely fashion. In the old BI4Dynamics version, the entire analytical database was deployed from an XMLA document, which was sent to the analysis services server by BI4Dynamics deployment wizard. The new architecture manages all dimensions and cubes into XML structured objects, which can be serialized into .NET Framework AMO objects. These objects are then deployed on the analysis services server by the new BI4Dynamics .NET client. All data warehouse and analytical database objects of the new BI4Dynamics data model, which conflate into the BI4Dynamics module (dimension or cube), are structured into XML files, which provide multi-versioning ability for each object. Deployment of the correct version of the object and the right deployment order is managed by the MEF framework implemented in the BI4Dynamics .NET client. The new BI4Dynamics data model brought benefits both in the trade and technical fields. The sales teams use a new tailored-based package sales model that allows adapting to the needs of the worldwide market. From the technical point of view, the new architecture provides a platform which can be easily expanded to other transactional data sources, such as Dynamics CRM and Dynamics AX.

    Item Type: Thesis (EngD thesis)
    Keywords: BI4Dynamics, data warehouse, analytical database, data model, architecture, Microsoft Dynamics NAV
    Number of Pages: 48
    Language of Content: Slovenian
    Mentor / Comentors:
    Name and SurnameIDFunction
    izr. prof. dr. Marko Bajec245Mentor
    Link to COBISS: http://www.cobiss.si/scripts/cobiss?command=search&base=50070&select=(ID=8026964)
    Institution: University of Ljubljana
    Department: Faculty of Computer and Information Science
    Item ID: 1214
    Date Deposited: 28 Oct 2010 12:59
    Last Modified: 13 Aug 2011 00:38
    URI: http://eprints.fri.uni-lj.si/id/eprint/1214

    Actions (login required)

    View Item