Design On Data Warehouse Performance Based On Model Driven

The characteristic of a software system is exhibited by its functional and non-functional properties such as performance, reliability and security, etc.. In many software systems, especially the large-scale one, non-functional properties such as data warehouse performance are forced to do.Performance property plays a vital role in guaranteeing the quality of data warehouse information model and system design. But currently the researches on data warehouse performance mostly focus on passively acquiring query performance target, usability target, and data quality target, at the same time, the design methods are organized disorderly in system architecture. In most project developments, performance design absolutely depends on the experience and ability of a software engineer, which possibly results in failure of data warehouse development. Therefore, researching systemically the data warehouse performance and giving the methods and guidance of performance requirements description, performance design and implementation are very important to successfully develop and apply the data warehouse system with high quality.In the work, performance design based on model driven is proposed, which is a method of describing the performance requirement and design using UML Profile. The meta-information of performance property is acquired by researching the basic characteristic of data warehouse performance. The design principles are abstracted from many optimization method of performance. The performance refinement K-Bs provide knowledge accumulation on performance design for most data warehouse software engineers.