The Design And Implement Of The Terminal Marketing Analysis System For Big Data

The development of the communication industry is changing from voice business to data traffic bussiness, single business to business interegation, mobile communication to mobile Internet. Terminal bussiness has been developed into an important link in “Cloud- Management- Terminal” as the strategic task of China Mobile, and it bears the burden of corporate transformation tasks. Terminal bussiness is not only the sources of revenue and resources of one telecom operator, but also an important leverage of the rapid development of value-added services and increasing data traffic and revenue. With the development of the terminal business, there will be hundreds of thousands of data directly related to the terminal business, and indirectly related data up to billions every day. The existing terminal management system can’t satisfy the need for huge amounts of data analysis.In this thesis, the technologis of storage, process and analysis related to big data were given firstly, and the technical route of the system research and development were also determined. Through the investigation and research on the existing service systems and marketing department, we analyzed the requirements of terminal management business and the existing shortcomings of the terminal management system, and obtained the function requirements of the system. Based on the results of needs analysis, the system architecture was presented in the level of the logical structure, application architecture and technical architecture, and the system boundaries with other business systems were determined. In this thesis, the development technologies and platform were introduced, and the big data infrastructure module and the definition and implementation of the data acquisition infterace were also defined. Terminal marketing analysis system is based on the cluster of Hadoop servers, and its data transmission interface with CRM system is defined as XML file. The business data of terminal marketing can be saved in HDFS as files, and the other data in other business system can be captured and stored using Sqoop. The terminal marketing data stored in HDFS can be analyzed by using Yarn framework. The interfaces of massive business data query, batch query, API programming and Web are implemented by HBase, Hive, Thrift and J2 EE technology, repectively. Finally, the testing and pilot application of terminal marketing analysis system is introduced.This system is based on the infrastructure of big data management, which implements the rapid deployment and extension of the Hadoop platform. The technologies related to Hadoop are used to implement the storage and analysis of data, and which are related to the businesses, such as procurement, storage, sales and other business of customized terminal in communication industry. It meets the demand of terminal information analysis of business department, and can help to exposed the potential risks and vulnerabilities, such as unpacking and inventory transshipment problems, as early as possible, protect the interests of the consumers and companies effectively. By using this system, the capability of terminal sales is improved, the terminal management processes are regulated, and the ability to control the whole lifecyle of the customized terminal are enhanced. The terminal marketing analysis system can boost the development of TD-LTE terminal business.