| Cloud computing provides a better solution for enterprises that provide network services.Because of its convenience and economic characteristics,more and more enterprises are putting the company's services on cloud servers to serve users.However,with the rapid development of the mobile Internet and the Internet of Things,more and more mobile devices are connected to the network.On the one hand due to the insufficient of cloud computing on the mobile Internet,on the other hand,the service requests of billions of mobile devices are handed over to the cloud computing to not only bring tremendous challenges to the computing power of cloud computing,but also cause congestion on the Internet backbone.In recent years,as the expansion of cloud computing from the center of the network to the edge of the network,a new computing model,fog computing,is proposed to solve the problem of inefficiency caused by centralized computing in cloud computing..Fog computing is a cluster of distributed servers located between mobile devices and cloud computing and close to mobile devices.Fog computing has certain computing,storage and communication capabilities relative to cloud computing.As an intermediate layer between mobile terminals and cloud services,fog computing can share some computing tasks of cloud computing and mobile terminals,reduce bandwidth delay,improve service quality of end users,and reduce backbone network bandwidth load and cloud computing center computing load.This paper studies the problem of storage optimization between cluster nodes in fog computing.Utilizing the storage space of the fog service node,the data service provider can pre-store the data that the user may need to access from the data center to the fog computing node closest to the user in advance.When the user needs these data,he first searches the neighboring service nodes.If the data is pre-accessed in the cluster,the user only needs to obtain the data from the service node and does not need to obtain it from the remote cloud center It can greatly reduce the access delay,reduce the load on the backbone network traffic.The main work of this paper is as follows:(1)Analyze the importance of fog computing node caching hotspot data to user service quality and network system efficiency.The hot data is placed in the fog node near the user,and the user can obtain the data through one hop distance of the wireless network,which greatly improves the delay of the user to obtain data.(2)Proposed a caching replacement algorithm based on Max-PSN,and calculated the cache effect value of cache data items in fog nodes.The larger the utility value of the data item in the fog node,the more it should be cached.When the space needs to be replaced,the data item whose utility value is the smallest is always replaced.(3)The centralized fog computing system is defined,and the operating mode and applicable scenarios of the system are given.Simulation experiments are carried out in a centralized fog computing system to verify the effectiveness of the three cache replacement algorithms that Max-PSN,LRU and LFU.Compared with the LFU and LRU algorithm,Max-PSN is 8.26%,2.87%,1.93%and 1.72%higher than the LFU algorithm and is 44.67%?11.12%?4.98%?3.75%higher than the LRU algorithm in one-hop hit rate,fog system hit rate,average response speed and bandwidth cost,respectively.(4)The distributed fog computing system is defined,and the operating mode and applicable scenarios of the system are given.Simulation experiments are carried out in a distributed fog computing system and the effects of three cache replacement algorithms that Max-PSN,LRU and LFU are compared.The experimental results show that Max-PSN is superior to LRU algorithm and LFU algorithm in one-hop hit rate,fog system hit rate,average response speed and bandwidth overhead. |
PDF Full Text Request