| As the decision-maker of management team in modern enterprises, especially in the world-class financial firms and international listed companies, the data warehouse has become indispensable tool, also including the board of directors, senior executives level manage. So we must build a unified, integrated, topic-oriented and storied historical data platform. It can integrate business information, form a unified view of the information in all over bank, and can more flexibly support a variety of analysis applications. And we can improve the quality of banking service and the market competitiveness, size the financial risk and improve the quality of bank assets, control the cost and improve the efficiency of banking operations and management, change the management model and keep the growing of business.Based on the Banking Data Warehouse development target, combined with the application demand for banking, balance system performance, storage, use and management of factors such as access, we build a data warehouse system. The system includes system architecture, logical architecture, system logic structure, data flow, internal logic-level, application mode, development environment. The specific component is primarily for the following several elements:1: the data sourceThe data source is as the source of ETL, and it give the raw data to EDW. The selection of the source data must have the accuracy, comprehensiveness, stability, continuity requirements, to reflect the long-term and comprehensive information and is thus in order to conduct an objective and comprehensive analysis and decision support.2:ETLETL is a data warehouse of data flows, the main link in the process of data warehouse data flow scheduling core. Data Warehouse ETL data acquisition program to obtain data, processed by the generic file transfers, transfer to the ETL server, from the ETL scheduling decompression, load data temporary data area, and then follow the rules of the source data mapping and data cleansing rules SDM data conversion to basic data area, generating aggregate data and application support interface, the ultimate market to provide data for the application.3: data servicesData Services is the enterprise-class data warehouse core part. The Data Services include temporary data area, basic data area, summary data area, and application data area.In the temporary data area, it has the cache of data source, ETL database, ETL processing temporary space. In the Basic data area, it stores 10 logical data model topics such as party subject, product subject, according to banking data warehouse subject. Summary of data area of the store is designed for quick queries, which contains all the frequently requested information. Application data areas, including meta-data database, data quality check database, the application support interface of data warehouse applications visit.4: meta-dataThe main objective of meta-data management is to provide a comprehensive guide to data resources. Metadata defines the data warehouse data model, source, extraction and transformation rules, etc. And the operation of entire data warehouse system are based on metadata. In the banking data warehouse, combined with business reality, the most appropriate meta-data management is "the Federal-style meta-data management framework".5: system monitoringSystem monitoring is to monitor, prompt diagnosis and reporting faults the network, system performance and operation condition, etc. Including data warehouse monitoring, MPP server monitoring, ETL Server Monitor, Web Server Monitor.Data warehouse system monitoring, including the operating status, the use of space and resource utilization of three aspects; Through MPP data warehouse server monitor, users can subscribe the needs of the monit events and the action. When a user subscribed event occurs, the corresponding monitoring system will have to implement user-defined actions; ETL server monitoring, monitoring of the ETL server operating systems, the ETL-related services, and log and monitor the implementation of the mandate; Web server monitor, is to monit operating system, websphere service status, log.6: system backup and recoveryThe main function and role backup and restore is to ensure full recovery the data at any point of time; least minimize data loss; to maximize the efficiency of the process of data backup.Way of data backup is usually a full backup and incremental backups. We only back up the previous changes of before. Recovery only needs a full backup and subsequent incremental backups. SDATA daily incremental backups can be used right the way back up the data area. Recovery operation, the first resumption of SDATA and then re-load PDATA area.7: project implementation managementData warehouse involved huge numbers of staff from different departments. So during the process of human resources management, it is necessary to strengthen the management of project members. They both have a reasonable role, have flexibility to adjust personnel to ensure that human resources in the warehouse of long-term development. It will ensure the project progress and quality. The project of data warehouse implementation are: project preparation stage, the project planning phase, requirements analysis phase, technical architecture design, source data analysis, data modeling, ETL development, application development.Banking data warehouse system implements centrally and standardized treat, as a cross-platform information. However, as the basis of data platform, it has responsibility to support the analytical applications of background data. Also, there are many co-ordination problems during the process. And it will still be some needs to continue improving. They will be the main objective of future work. |
PDF Full Text Request