Design And Implementation Of Massive Audio File Storage System Based On HADOOP

With the rapid development of communication service industry, each enterprise business type and the service classes are becoming diversified, interactive user and enterprise customer service number is growing rapidly. Many of the company’s customer service is also increasing pressure, the company in order to guarantee the normal work and service personnel, to prevent malicious telephone harassment, and in order to ensure that the personnel of the service business in accordance with the requirements of customers, usually on the customer service call recording preservation. The recording document large quantity, but also in the rapid dynamic growth for the company, but they later data analysis, used to improve employee service quality management is very important, usually long term preservation. The traditional storage method of processing the data of low efficiency, heavy workload, and the cost is high. How safe and effective and low-cost storage of these data is a problem that should be solved urgently.This paper mainly studies how to use a variety of storage equipment, building a distributed storage system, in order to solve the problem of massive recording file storage. This paper analyzes the present situation of domestic cloud storage technology development, through the research of distributed file system technology, combined with the storage requirements for the massive recording file, the design and implementation of massive audio file storage system based on HADOOP. The system is based on the development of a mass storage recording file storage system based on HDFS, the system is composed of client, message queuing and service storage end. The client is mainly to provide a friendly user interface, the message queue is a link between the client and server, server implementations of distributed storage and respond to client request all kinds of. The system is a typical production consumer model, the client request message to the queue, the server processes the message and response. The system server is dynamically extensible, new server program, service program as long as listening to the specified message queue can receive and process the client’s request. The following describes the demand to the system implementation steps:First, through the analysis of the status of enterprise massive recording file storage, considering deficiencies of traditional audio file storage mode, this paper puts forward a distributed storage technology research, massive audio file system based on HADOOP;Second, combining the current situation and the requirement of the massive recording file storage, the massive recording file storage system is designed in detail, mainly includes the user authority design, identification of recording file upload and download design, the recording file sharing design;Third, the environment of distributed file system structures is introduced in detail, and the realization of the massive recording file storage system, including user access identification function realization, the recording file upload and download function realization, the recording file sharing function realization. Considering the portability and security of the system, using Java language to development. The system is divided into client, server, database and distributed storage system. The system interface is simple and clean, convenient for users to use;Fourth, the system of security, file transfer efficiency and the function of the system to do the test, in the recording file storage method, compared the system with the traditional system, safety performance, data transmission efficiency is better also has greatly improved, the data storage costs are lower.