The Research On Pre-loading Policy Via Analyzing User Switching Behavior In IP-based TV Systems

With the rapid development of mobile communication technology and wide application of mobile terminal,IP-based TV systems,which rich in performance,swarm into people daily life and are widely used.Compared with the traditional broadcast TV systems,IP-based TV systems give users more opportunities to experience channel switching delay on channel switching.In this paper,we construct a state transfer model to describe user watch behavior in IP-based TV systems and to analyze the characteristic of users' switching behavior and channel switching.According to the users' channel switching characteristics,we propose two pre-loading policy for achieving close-to-zero channel switching delay at the price of extra download bandwidth.The one is about how to find the right person to pre-loading,the other is about how to select channel used for preload.Firstly we have shown the research on network TV users' viewing behavior and pre-loading algorithm at home and abroad.Also the technical field related to the network television system had been researched through an in-depth discussion and research which analyzed the composition and causes of channel switching delay.In the third chapter,a network TV user view state transfer model describing users' switching action as transfer between states,will be established,which makes the arbitrary user viewing process can be described as a loop of the nodes and edges.From simultaneous analysis of network TV system Banacast users' viewing data,two characteristics of user switching behavior have been found: firstly,the probability that a user does not switch channel is far greater than the probability of channel switching;and secondly,after the channel switching,users are more likely to watch for a long time and the users' view of watching time basically conforms to the power-law distribution.In the third chapter,the entropy of switching to a channel and the entropy of switching out a channel will be proposed to analyze the system channel switching mode.It is found that the channel switching mode is sporadic in the early morning and is centralized in the evening time.Finally,in the fourth chapter,according to the characteristics of network TV user switching time and switching mode,we proposed a pre-loading user selection algorithm and a pre-loading channel selection algorithm which can pre-load channel data adapt to the bandwidth.In short,the analysis on time characteristic and channel switching mode of the main network TV user switching behavior fill the domestic blank of research on network TV user switching behavior.And the pre-loading algorithm in this paper can adjust itself to adaptive pre-loading channel according to the change rate of switched systems.Finally,not only can it effectively reduce the bandwidth consumption on pre-loading and improve the utilization rate of pre-loading data,but also do finally achieve zero switching delay,providing a better user experience.