Study On Strategy Of Media Data Storage And Propagation For IPTV System

With the trend of the integration of the telecom network, the Internet and television network, the user's growing demand for Internet applications is becoming quite diversity. IPTV, as one of the most popular multimedia applications in recent years, comes into being. However, there are lots of problems in IPTV system such as the network performance, media data storage and distribution and so on; Millions of users and massive multi-media data bring lots of difficult problems to the system. So,how to improve the quality of services and the experience of user as much as possible is a matter of life and death for the emerging industry of IPTV.To IPTV system, the strategy of media data storage and distribution is very important to the performance and service quality. This thesis mainly studies on the strategy of media data storage and distribution.First of all, this thesis implements the CDN based media data storage of IPTV system. CDN changes the system from a single central structure to a distributed one, which greatly reduces the pressure of backbone network besides ensuring high-quality.Secondly, this thesis provides a new model for media storage and distribution based on the statistical analysis of field data. By the analysis, we discover that users'visiting habit of VOD is similar to obey Zipf distribution. Therefore, we apply Zipf law to the analysis of users'behavior, and build a prediction model based on this. By tracking users'behavior, the new model predicts which contents are most likely to be visited frequently in the future,and pushes these contents onto the nearnest media station and deletes the cold contents from cache at the same time, which ensures high hit rate besides the realization of dynamic storage.The new solution reduces the waste of media storage space and improves the experience of user.In addition, we try to improve the storage and distribution system, and implement a scheme of slice and hierarchical storage for media data, in order to resolve the problem of load imbalance and reduce response time of time-shift TV.