Research And Practice In The Shanghai Mobile Detail Billling System Based On Cloud Computing Platform

IT Business system in China Mobile company at Shanghai is a very huge system and it can support,illions of users. The Billing System in it must manage billions of bill every day.The system aboutdetail billing managment is a part of the Billing System. It could support about 8,000,000 users in maximum while the active users in Shanghai Mobile is only 6,982,000 at the early stage of construction.However in the face of the current Shanghai mobile users exceeded 20,000,000, the system is facing crisis. With the the billing detail recent growing,the performance of the server and the storage needs to be expanded in order to let the customers check their telephone statement more efficiently.Moreover, the current system still exist the following problems:the high cost of system expansion, the internal demand for query and statistical become more and more complex, the response of the demand become slowly,the efficiency and the benefit ratio must reach a higher level,etc.Therefore,we will introduce cloud-computing technology to rebuild the system.HADOOP is an open source JAVA framework, and it is commonly used in distributed computing. It has a lot of the advantages such as high dependability, good performance, flexibility.Moreover it is widely used in many big data and cloud computing area. It may be supported by the MapReduce programming examples for calculation, and it can store the data through a HDFS(Hadoop Distributed File System).It is recently applied in many large websites and be called “the most widely used cloud”.The article first introduces the Hadoop Technology, the development and the current situation.Then it introduces two main parts of Hadoop: HBASE and HDFS. A detail query system is established by specific research and practice in the Shanghai Mobile's Detail Billing system based on hadoop platform. It can realize a high efficiency in data querying and loading with the storage of detail billing data in HBASE by the efficient way of MapReduce. Then the system is installed, commissioning, including functional testing and performance testing and compared to the old system. The test results show that, the system can provide more efficient query service.Finally, we put forward some perfect content about the prospects to this system for the future.