| As people entering the era of big data,data assets have turned the core competitiveness of enterprises.How to build an ecosystem of big data around the large data assets of telecom enterprises and a business model with the orientation of user experience,has become a significant task faced by telecom enterprises.The principle of the construction of telecom business data analysis platform is to build a perfect high-performance open platform.Stability of the big data analysis platform is the most important,while the key points include performance of components and upgrading of some technique.It is necessary to provide an environment of big data application development,the Develop Studio,to make sure that partners in big data could directly develop,release and run applications on this platform.Meanwhile,basic framework of the platform should be realized on cloud.Based on the above,we have designed the framework of the telecom business data analysis platform based on Hadoop in this paper and realized some functional modules.First,this paper briefly introduces the Hadoop architecture and general development technology,including the history and status quo of Hadoop and basic system architecture.Then it carries on the design and implementation of the telecom operators’ big data platform based on Hadoop,making some designs of the platform data and the technical architecture,as well as designs of some information integration module,data assets module and open ability module in this architecture.After that,it comes to design the application characteristics of Hadoop cluster in this special system and make some study on the enhanced Hadoop framework,in which designs of enhanced functions in the system,such as the HDFS,MapReduce,HBase,MPP are discussed.Next,this paper describes the realization of the resource management module in the platform,makes an implementation of the operational control module and especially fulfils the management of metadata and data security in detail.At last,summaries and prospects of the telecom business data analysis platform are discussed. |