Collaborative Caching Strategy For Mobile Edge Video Transport

With the advent of 5G and the interconnection of all things,the computing,storage and transmission costs of cloud computing increase exponentially,together with its inherent large latency and vulnerability,making people have to look for new computing models.Different from cloud computing,in edge computing,Mobile Edge Computing(MEC)servers provide computing and storage resources,and MEC servers cooperate with each other to expand service capacity and service scope.At the same time,because MEC servers are closer to users,they can better meet users’ requirements for response delay.These characteristics make edge computing the best way to solve existing problems.Due to the large number of MEC servers in the edge network,how to efficiently collaborate between MEC servers and seamlessly connect with Content Delivery Networks(CDN)will become a key issue for mobile edge computing.In addition,most of the existing edge pre-caching schemes do not take into account user preferences,and their accuracy needs to be improved.Therefore,recommended techniques can be incorporated into the pre-caching strategy to enhance the user experience.Combining the characteristics of video caching and edge computing,this paper presents a video Collaborative Caching and Replacement(CCR)scheme based on cache benefit and replacement cost,and also presents a Proactive Caching Strategy Considering User Preferences(PCSCUP).This paper contains the following aspects of research.1)This paper presents a multi-MEC node collaboration architecture which seamlessly integrates with CDN.This architecture is composed of MEC servers,mobile devices and cloud servers.Neighboring multiple MEC servers make up a collaboration cache domain.Cooperative caching enhances the cooperation between MEC nodes and improves the overall efficiency of the MEC collaboration cache domain.The architecture can accommodate a large number of MEC nodes,providing ideas for the collaboration of large-scale edge networks.Based on this architecture,this paper simulates a complete set of edge cache experimental environment.2)This paper proposes a video cache replacement strategy based on cache revenue and replacement cost,which comprehensively considers the positive revenue of cached video and the negative revenue of replacement video.In terms of caching,considering queue delay,send delay and video popularity to calculate cache revenue,and then select the best cache node in the MEC collaborative cache domain.In terms of replacement,this paper considers the cost of reacquiring videos and the popularity of videos,selects the video with the lowest replacement cost,and executes the cache replacement decision.The experimental results show that the cache replacement strategy proposed in this paper can effectively improve the performance in terms of cache hit rate,delay and transmission load.3)In order to improve the accuracy of pre-caching,this paper proposes an proactive caching strategy that considers user preferences,and proactively caches when there is a big difference between the preferences of newly launched users and the group preferences of the MEC where they are located.This strategy can adaptively adjust the size of the proactive cache space according to the cache hit rate,making the use of the cache space more intelligent.The experimental results show that the proactive caching strategy performs much better than the non-proactive caching strategy,and can effectively improve the performance of QoS,such as cache hit rate,etc.The strategy has good adaptability to users with different preference distributions and videos with different popularity distributions.