Research On The Qos-based Scheduling Between Updates And Queries In Real-time Data Warehouse

As the development of the information construction in enterprises and organizations, they have collected a lot of data, which provide basis for decision making to use data warehousing technology, online analytical processing (OLAP) technologies and data mining technologies to make a quantitative decision. Data warehouse system has become an indispensable assistant tool in production and operation decision-making. However, the traditional data warehouse is generally cyclical update, and update cycle is a day or a week. So the data for the decision don't include the latest information. Accordingly, the strategic decision based on traditional data warehouse can not meet the real-time requirements. Therefore, it turns to the real-time data warehouse (RTDWH).It is important to capture the changed data in time, so that the data warehouse can be updated with the data. Queries on data warehouse is online, consequently it may bring the collision between updates and queries. This article is mainly engaged in solving the two problems above.To the problem of change data capturing, we use CDC tool in Oracle to accomplish capturing changed data. In the article, we offer a method of changed data capturing with CDC tool, and we also give corresponding flow design. At the same time, we prove the preciseness of the scheme. On the problem of updates and queries scheduling, we combine the internal and external mechanism with QoS to guarantee data freshness and query speed. External mechanism is the arithmetic which users give the need for query speed and data freshness, while internal mechanism is the arithmetic which executes update forcedly when the system update is abnormal. As the results reflect, the scheduling we offered does well in solving the confliction between updates and queries.