Research On Caching Strategy In P2P Streaming Media

The introduction of P2P streaming media technology has impacted the traditional video transmission technology.This novel streaming media transmission method has been widely used in streaming media systems.The caching strategy of the P2P streaming media system needs to accurately evaluate the hot video in order to improve the hit rate of the video segment to a high degree and reduce the response delay.This is directly related to the user experience and the performance of streaming media.In addition,the node selection technology in the P2P streaming media system can not only reduce the pressure on the server,but also improve the service quality of the service nodes.Therefore,the cache strategy proposed in this paper starts from two aspects of node selection and video prediction.The specific research work is as follows(1)Because the selection of cache nodes will affect the cache efficiency,in order to solve the problem of inaccurate node selection in P2P streaming media cache.This paper proposes a new node selection algorithm using the greedy algorithm(GA-PSA),which comprehensively considers the uplink and downlink bandwidth of the node,the online time of the node,the distance between the nodes,and the ability of the node to serve.Therefore,an efficient cooperative node is selected under a highly dynamic P2P network topology.In this paper,a global optimal problem is transformed into multiple local optimal problems and solved quickly.Simulation experiments show that this technology can improve the transmission delay and throughput of the system,and effectively improve the overall performance of the system(2)Because the popularity of some video segments that have been broadcast for a period of time does not change much,using frequently updated models cannot accurately describe the popularity of video segments.Therefore,this paper proposes a cache replacement algorithm based on association rule(CRA-AR)for video segment prediction.This strategy is more focused on storing video segments with higher popularity.This strategy mines the user's historical access records through association rules,predicts multiple segments to be cached,and then filters based on popularity to obtain the final segment to be cached.Simulation experiments show that the technology has better performance in terms of hit rate and response time(3)The popularity of newly released video segments has not formed a stable trend,traditional statistical methods cannot reflect the changes in popularity in time.In response to this problem,this paper also proposes a prediction algorithm based on the Modified Markov Prediction Model(MMPM).This strategy can be run with a small number of historical user access records.First,a Markov model is established based on user historical access records,and a state transition matrix is obtained.Re-integrate the exponential movement weighting model so that the state transition matrix contains past information.The state transition matrix is modified in real time,so that the strategy can predict the segments to be cached more accurately.Simulation experiments show that achieving dynamic prediction improves the hit rate and response speed.The effectiveness,accuracy and speed of the algorithm are verified.Considering the above two aspects,the algorithm proposed in this paper not only has the advantages of robustness,scalability,high cost performance,and load balancing,but also extends the application scope and practicability of this technology in the field of streaming media.It is of great significance to improve the playback quality of streaming media.