Design And Implementation Of Metadata Management Solutions For The Massive Data Analysis Platform

With the rapid development of information technology, the fast growth of enterprise data and increasingly complex business logic are beyond the scope of traditional Business Intelligence (BI) system’s solution ability. So the massive data analysis platform is proposed combined with BI and PaaS which is one service model of cloud computing. However, the platform still has some problems, especially in the aspect of metadata. It is lack of complete metadata system and effective metadata management system. And this will seriously affect the flexibility, the scalability and many other properties of the platform.The objective of this paper is to design and implement the metadata management solutions of the massive data analysis platform. Through metadata management, we could achieve configurability, flexibility and scalability of the platform.First of all, this paper designs the platform metadata system according to the three categories of manage metadata, technical metadata and business metadata. And it makes a detailed classification of the technical metadata in accordance with the platform data processing.Secondly, this paper puts forward a metadata management framework and designs" manage meta-model and technical meta-model using UML class diagrams mainly. Then, it develops model mapping rules to map these models to tables and attributes in the database.Thirdly, this paper designs the platform metadata management solutions, including basic maintenance of metadata and the management of the relationships between them. This paper utilizes the technology of JSP+SSH (Struts2, Spring, Hibernate) to implement this project. Then, it uses the temperature application as an example to analyse the metadata management processes in order to illustrate the superiority and reasonable of metadata management solutions.At last, this paper summarizes the main work and deficiency of the papers, and then puts forward the future work.