Design And Implementation Of ETL Tool In Telecommunication Statistics And Analysis Platform

Value-added services have become the major income source of telecommunication operators. In the market driven by users' requirements, analyzing the user requirements and digging out user features become the focal point of new service development. And Telecommunication Statistics and Analysis Platform is designed to meet such demands.CSP (Colorful Ringback Tone Statistics Platform) is a statistic platform independent from CRBT (Colorful Ringback Tone) platform. CSP collects operation data from CRBT platform and perform data analysis work to produce statistics to assist marketing strategy.SAP (Service Analysis Platform) collects service operation data of IN (Intelligent Network), analyzes and evaluates the status of service, and provides automation assistance for the IN maintenance personnel.Both CSP and SAP are typical telecommunication statistics and analysis platforms, whose core are data of huge amount and diverse types. To integrate these data into the analysis system, specific tools are needed, which is ETL (Data Extract, Data Transform, Data Load). ETL extracts data from various environments, clean and transform the original data and load the clean data into destination database, on the basis of which, services of higher levels could perform their work efficiently.Telecommunication Statistics and Analysis Platform has a number of specific requirements for the ETL process: The ETL should have the ability to deal with large variety of data types and huge amount of data; The ETL should adapt itself to the change of metadata, and should support the real-time data collection. These requirements could be reflected clearly in the CSP and SAP system. To meet these requirements, detailed analysis is needed, on the basis of which specific solution for each requirement could be designed.In the process of the system design and implementation, large amount of experiences are accumulated, and features of the ETL tool in telecommunication statistics and analysis platform are summarized, and some guiding principles could also be deduced. By combining these experiences and principles, a general ETL architecture that could be applied in the telecommunication statistics and analysis platform could be generated. With a flexible structure and ideal performance, this architecture could not only meet the existing requirements, but also could evolve together with the telecommunication statistics and analysis platform to become the new generation ETL tool.This thesis focuses on the design of the ETL in the Telecommunication Statistics and Analysis Platform: First of all, an outline of CSP and SAP is given, in which ETL technology was applied; then the relevant background knowledge is introduced, including Data Warehouse and OLAP (Online Analysis Process). In Chapter 2, key issues encountered during the development are analyzed, and methods for each issue are described. Based on this, features of the ETL tool in Telecommunications Statistics and Analysis System are summarized. In Chapter 3, the design and implementation for General ETL architecture are proposed. The last chapter analyzed the future development of Telecommunication Statistics and Analysis Platform and the future application for General ETL Architecture.