Design And Implement Of Collaborative Hierarchical Caching System Based On Edge Computing

Driven by a new round of scientific and technological revolution,the trend of digitization,intelligence and the Internet of everything is becoming more and more obvious.The explosive growth of network traffic and the number of terminals brings a huge burden on computing capacity and network transmission of cloud center.How to reduce the network bandwidth consumption,decrease the user delay and improve the quality of service is the top priority for the future development of the Internet.Therefore,edge computing which can provide users with lightweight cloud services at the edge of the network,and caching technology that caches popular content to respond to user's requests more quickly,begin to be highly concerned.In order to solve the problem of excessive user delay in cloud computing model,this paper combines the characteristics of edge computing and caching technology,studies the design and implementation of caching system based on edge computing.The main contributions are as follows:First of all,this paper optimizes the edge computing framework and proposes a framework optimization scheme which is more suitable for caching.Based on the resource differences of edge nodes and local area network topology,the edge computing framework is optimized into a four-layer structure of "cloud-scheduling layer-access layer-terminal".The edge node in the scheduling layer is called the regional edge server(RES),which is responsible for managing the edge cluster,communicating with other clusters and caching the popular content within the service scope of the cluster;the edge node in the access layer is called the local edge node(LEN),which is responsible for caching the popular content within the service scope of the node.This optimization not only provides a basis for collaboration between different clusters,but also realizes the hierarchical placement of cache content according to popularity,which reduces the cache redundancy of the whole cluster.Secondly,based on load average,disk I/O idle rate and bandwidth idle rate,this paper puts forward the concept of loadable factor,which is used to indicate the ability of the server to respond to user's requests without overloading itself.At the same time,this factor can be used as an important parameter to select collaborative clusters.Then,combining with Newton's law of cooling,the cumulative number of requests and the recent trend of a certain content being requested,this paper proposes a method to calculate the popularity,and uses this method to design the cluster's cache placement and replacement strategy to improve the cache hit rate and space utilization of the whole cluster.Based on the above research,this paper designs and implements a collaborative hierarchical caching system based on edge computing.The system is divided into storage side and control side.The storage side is responsible for the storage of cache content,which is composed of the cache storage unit of RES and all LEN in the cluster;the control side is deployed in RES and responsible for controlling the whole cache system,which is mainly composed of information management module,node monitoring module,cluster's cache placement module,cluster collaborative caching module,cache reading module and cache cleaning module.The information management module is responsible for managing the cache and device information of the system;the cluster's cache placement module selects popular content and stores it according to the popularity calculation method proposed in this paper;The cache cleaning module is used to clean and replace the cache content of a node when the storage space of this node is insufficient;the cluster collaborative cache module is used to find and use the appropriate cluster for cache collaborative when the loadable factor of RES in the cluster is too low.Finally,this paper sets up a real test environment to access the resources in the web server concurrently,and compares the cache hit rate,average delay and other indicators with the classical cache strategy in the cache service.The test proves that the collaborative hierarchical caching system based on edge computing can effectively decrease user delay,reduce the cache redundancy,improve the space utilization of caching system,and prevent the serious overload of RES.