| The key of the successful enterprises is a running good data center. Now, they hope Provide business department in a professional, standard and efficient IT services by constructing data center in succession. The determinants of data center function well are people, processes and tools, quantify these three aspects of KPI is the prerequisite, especially for the analysis and enhancing the operating efficiency of people, processes and tool. Throughout the statements management status of domestic data center, the problems mainly include:1,The data view were dispersed and no uniform platform provided dynamic data;2,The mass data not cleaned and processed,can't be used for policy makers.Business Intelligence (BI) technology through that cleaning, sorting and transformation of mass data aid decision making, even generate business decisions automatically to improve business data analysis efficiently, so BI is widely used. The paper combining with the actual conditions of enterprises through extracting, transition and loading, finished concentrated statements platform of the data center, through the data on-line analytical processing and data mining, find out useful information for management decision makers.The paper firstly discusses the state of data center indicator, then declares the necessity of designing and completing centralized and unified statements platform based on BI of data center. On this basis, combining with decision tree model, As the user segmentation module for example, a detailed description data extraction, data cleaning and the final decision tree display model, research and implementation based on BI data center report systemThe report system of data center through determining demand of data centers concentrated statements, Including functional requirements and non-functional requirements, designs BI framework of report system, and uses the three-layer structure model, makes display logical, business logic and data logical separation, optimizes the design of BI frame. The base of BI is data warehouse based modeling, this paper first determine particle size of data warehouse, combining data center current portfolio and practical business, to determine the size of the data warehouse and its model structure. The basis of BI implementation is ETL process design, combining data center current portfolio and the practical operation, complete the ETL process design, to ensure the accuracy of the data warehouse, the error table is established, they are used to check 1 each transaction data of data warehouse. At the end of the paper, statements data achievements is showed, including histogram, instrumental etc. |
