Research On 360° Video Caching Strategy Based On SVC Coding In ICN

In recent years,with the rapid development of virtual reality(VR)and augmented reality(AR),360-degree panoramic video has grown up to become a popular Internet service.Due to 360-degree panoramic video's high bandwidth and low delay requirements,the transmission of 360-degree video streams on the Internet has problems such as limited bandwidth and poor user viewing experience.Although using tiling technology and Scalable Video Coding(SVC)to transmit 360-degree video can reduce bandwidth consumption.However,due to network latency,some problems still occur such as long acquisition time and excessive playback pauses during fetching 360-degree video streams from a remote content server.Therefore,the transmission of 360-degree video is still challenging.Caching technology is a powerful solution.Properly deploying caching between users and content servers can significantly reduce round-trip latency and bandwidth consumption.However,the unicast communication mode of the TCP/IP network fundamentally limits the development of caching technology,which is particularly unfavorable for large-scale 360-degree video distribution.The ICN network is a revolutionary solution that can fundamentally solve the problem of the TCP/IP architecture and has huge advantages for large-scale video distribution.Firstly,this thesis makes full use of the characteristics of ICN network ubiquitous caching,and proposes a SVC-oriented Fo V-aware Probabilistic Caching(SFPC)strategy for 360-degree video.the SFPC policy selects the node in the transport path that has the largest available cache capacity to cache the video chunk during the newtwork initialization phase.when the initialization phase is over,the SFPC strategy caches low-level video chunks in Fo V on nodes closer to the user,and caches high-level video chunks on nodes closer to the server.SFPC also reduces cache redundancy by strictly controlling the proportion of cached copies of the same content block on the request path.Secondly,this thesis proposes a cache replacement strategy named "Size-Popularity-Layer-Fo V"(SPLF)and proposes a new video chunk popularity calculation formula based on the arrival time and access frequency of each video chunk.On this basis,according to the Fo V,popularity and other information of the chunk,SPLF calculates a cache value for each video chunk,and use it to measure the possibility of the video chunk being accessed at a future moment.When the cache is full,the video chunk with the smallest cache value is expelled from the cache.Finally,this thesis conducts simulation experiments on the head movement data set of real users in the ccn SIM module based on OMNet++.The simulation results show that compared with various baseline algorithms,the SFPC decision strategy and the SPLF replacement strategy can not only improve the hit rate,but also reduce the playback pause and rebuffering time.