Research On Key Technologies Of Content Caching And Edge Computing In Mobile Internet

The mobile Internet has a profound impact on the current social economy and industrial development.As the key development direction of the "Internet+" strategy,the mobile Internet has developed rapidly in recent years.It has become one of the essential pillar industries of national development and social progress.However,the current development of mobile Internet still faces many challenges:(1)the coexistence of resource competition and waste.Lack of resources in the mobile environment,severe competition in hot spots,and difficult access to the network lead to a severe waste of resources;(2)dynamic differentiation of user needs.The traditional static service mechanism can not adapt to the dynamic and personalized user demand because of mobile users' dynamic location and the difference in request behavior.(3)Mobile network resource distribution is dynamic discrete.The resources of different nodes lack coordination.It is not easy to be effectively managed and utilized;(4)the network environment is heterogeneous and complex.Different communication modes are independent of each other,multi-form network fusion is limited,and cross-network coordination is complex.The resource status constrains mobile Internet performance,seriously affects mobile network performance,and it is not easy to provide users with a high-quality service experience.Many researchers have done a lot of work on content caching and edge computing for mobile Internet to solve this problem.However,the current content caching strategy is relatively static.Performance depends heavily on the popularity of the content and the lack of users' personalized needs.Simultaneously,the mobile environment's cached deployment brings a significant security risk to the user's data privacy security.On the other hand,although edge computing is used in many fields,the lack of an effective solution to integrate the fragmented network's computing resources makes it challenging to improve the overall performance and provide high-quality,high-stability,low-latency mobile internet services.In this paper,two key Mobile Internet technologies,called content caching and edge computing are studied:coded caching,cache prefetching,cloud-edge-end computing integration,and joint optimization of computing and transmission.Firstly,an edge coded caching strategy for mobile networks is presented.The state evolution model based on the dynamic model is constructed to realize the cognition and prediction of network cache supply and demand.A feature-learning-based coded content selection framework and a privacy-aware coded caching algorithm are proposed to improve the efficiency of cache resource utilization and user privacy's security performance in mobile networks.This paper proposes a privacy-oriented mobile content cache prefetching mechanism,which includes distributed learning-based user interest awareness,multi-objective optimization-based dynamic online content caching,and differential privacy-based data privacy protection.It realizes the integrated scheme design from user's personalized cognition,active content cache to privacy and security.Then,a real-time computing and transmission optimization framework based on cloud-edge-crowd integration is proposed.A novel augmented queue structure is proposed to model the computing workload of nodes and congestion of links,the formal representation of the joint optimization problem of transmission and computing resources is realized.A distributed optimization algorithm based on the Nesterov acceleration gradient is designed.The theoretical performance indexes of stability,queue length,optimality,and convergence rate are given.It effectively integrates the computing resources of crowd nodes and provides good scalability for system capacity.Finally,based on multi-agent reinforcement learning,a joint optimization method for computing transmission,and an augmented graph model is proposed.The abstract joint optimization problem is transformed into an intuitive network routing problem,and a networked multi-agent reinforcement learning method is designed.The performance of collaborative computing and data transmission is improved.In this paper,we take mobile streaming as an example and study content caching and edge calculation in mobile network.The research methods include model building,problem formulation,distributed optimization algorithm design,algorithm theoretical performance analysis,algorithm numerical performance analysis,prototype system,and prototype experiment evaluation.The achievement of this paper has specific reference significance to the future construction of mobile Internet and the development of advance multimedia service in our country.