Research On Cache Placement Algorithms For Video Content Distribution In ICN

With the rapid development of Internet applications,the network traffic of traditional long video applications has greatly increased.In addition,with the rise of a number of video applications represented by short video applications and the continuous improvement of users' experience and demand for quality of service,communication network operation is facing great pressure.Studying video content distribution technology is of great significance to improve user experience and reduce network load.In order to adapt to the transformation of Internet applications from end-to-end communication mode to mass content distribution mode,and to solve the key problems of efficient content distribution from the network architecture level,the academic circles have proposed the Information Centric Networking(ICN)architecture in recent years.ICN has technical characteristics of content-oriented naming,routing and caching based on content names,etc.Compared with the existing IP network architecture,ICN has great advantages in content distribution,which can greatly improve the efficiency of content distribution,reduce the network load and reduce the delay for users to obtain content.As a commonly deployed resource in ICN nodes,cache provides users with closer content replicas,which can effectively improve content acquisition efficiency and resource utilization and reduce network load.As a specific al-gorithm to guide the deployment and operation of content in the caches,cache algorithm directly affects cache performance.Among many cache algorithms,cache placement algorithm has the most obvious influence on cache perfor-mance,and it is of great significance to improve the performance of content video distribution services.Therefore,this dissertation focuses on the research of cache placement algorithms in ICN.Based on the above background,the doctoral dissertation mainly studies the cache placement algorithm of ICN for video content distribution.Aiming at the problems existing in the current cache placement algorithm in various subdivided video content distribution scenarios,the doctoral dissertation car-ried out the research on cache placement algorithms.The main contents and contributions are as follows:(1)An on-path cache placement algorithm based on user interest prefer-ence is proposed.In short video scenes,the local popularity distribution at different nodes is quite different,and this difference has a more obvious impact on cache performance under irregular topology.In order to better grasp the characteristics of request distribution and improve the performance of cache placement algorithm,the user interest preference factor is introduced into the cache placement algorithm.Firstly,multi-dimensional content attributes,user interest preferences and their acquisition methods are proposed,and then the in-terest matching degree of cache nodes for certain types of content is proposed.Then,according to the index of interest matching degree,local popularity and cache space size of cache nodes,an on-path cache placement algorithm is de-signed.Simulation results show that the algorithm proposed in this chapter can greatly reduce the number of replacements,and improve the cache performance such as cache hit rate and hop count reduction rate.(2)An off-path cache placement algorithm based on the nearest replica routing strategy is proposed.For the scene with existing contents of long video,the current cache placement algorithm has the problem of low cache utilization due to limited routing range.In order to improve cache utilization,the combi-nation of cooperative cache placement algorithm and off-path routing strategy under arbitrary topology is studied.A cache placement algorithm based on the nearest replica routing strategy is proposed.The algorithm is divided into two sub-algorithms,namely,a replica allocation algorithm based on auction theory and a replica placement algorithm based on matching theory.The heuristic al-gorithm is used to implement a off-path cache placement algorithm with low complexity and excellent performance.Simulation results show that the algo-rithm proposed in this chapter greatly improves the performance in terms of load balance,average hop count and cache hit ratio of cache nodes.(3)An off-path cache placement algorithm based on popularity predic-tion is proposed.For the scene with new contents of long videos,the current off-path cache placement algorithm cannot place new content quickly,which leads to the inability to better cope with the possible instantaneous large num-ber of requests.Firstly,a new content-oriented popularity prediction algorithm is proposed,which can accurately predict the content popularity of new con-tent after a specific time under the condition of limited available information,so as to provide a cache basis for the cache placement algorithm.Then,a new content-oriented off-path cache placement algorithm is proposed.By solving the joint optimization problem of cache placement and replacement,the cache placement algorithm for a single new content and the replacement algorithm for existing contents can be quickly obtained with low complexity.Simulation results show that the popularity prediction algorithm and the cache placement algorithm proposed in this chapter are superior to the existing algorithms in prediction accuracy and cache performances.