Design And Implementation Of Online Video Recommendation System Based On Cloud Computing

Scientific Technology like Cloud Computing, Web2.0, and Recommendation Engine has been developed in an amazing rate. Development of technology is changing the way in which people are using to interact with the Internet; information technology has thus been improved. People’s demand of interaction between users and Internet has been improved when the Internet is emerging into people’s life. Recommendation System is kind of way to reflect people’s increasing demands. All the popular websites contain the recommendation system because their users want to obtain the useful information due to their preferences. But limited by the traditional technologies, recommendation system could rarely taking use of advantages of distributed systems. Redundant information in the Structured Query Language database and loss of effective distributed framework are hindering progresses of distributed system’s development. Such situation brings us the new Hadoop, No Structured Query Language database (NoSQL) and fresh Web development framework. NoSQL supports the distributed computing and handles well about massive data. Hadoop as a distributed framework, which offers the distributed and pseudo distributed developing strategies are laying a solid foundation of developing distributed system.This thesis takes use of a typical NoSQL database which names MongoDB, combining with Python Django-a Web2.0developing framework and Hadoop-a distributed development framework, design and finishes the coding of a distributed video recommendation system. By making the summary design, detail design and definition of data modules, the author decided developing tools, design patterns and recommendation algorithm. Design of distribution system, design and coding of the recommendation algorithm based on users’preference tags and finishing design of views module and templates are major works of this project. At last, the system was tested. Test plan of this system are consist of three parts: basic function test, compatibility test and maintainability test. By designing the sanity test cases, the system is proved to be stable and useful.Based on the design above, this thesis brings about the video recommendation system. Recommendation Engine, which is configured on the distributed development framework, analyzes users’history tags and recommends videos users may like. Other small functions have also been coded, such as users’registration, login and add new videos to the system.