Research On Fog Computing Based Named Data Networking Performance Optimization

The existing IP address-based network architecture was invented in the early 19 th century.With the proliferation of network data traffic and network devices connected to the network,this architecture can no more meet the needs of people.Information Centric Networks as the mean trend of next generation network attracts the attention of the public.Compared with the traditional IP network architecture,it uses name to identify the data,and therefore no restrictions on the number of IP addresses,which meet the demand of numerous devices connected to the network.Information centric network's characteristics of on-path caching make the distribution and re-use of data become possible,which effectively reduces the delay of sending and receiving data.Moreover,named data networking(NDN)as a mainstream implementation of information centric network has aroused widespread concern in academia domain.Although NDN has numerous advantages over traditional IP networks,it still has many problems.First of all,the limited storage resources of network equipment makes the advantages of on-path caching cannot be fully realized,compared to the amount of data in the entire network,the data hit rate of network equipment is very low,which has only marginal improvement of the network performance.Secondly,limited computing resources of network devices are not sufficient for high-complexity network control,including routing,complex caching strategies,and data integrity verification.Thirdly,the NDN proposed caching the data along all the path,which will lead to data redundancy and waste of resources.At the meantime,the performance improvement of network is ignorable.Finally,the mobility problem of end users in the network.For NDN,user movement will lead to increased packet loss rate,network congestion and frequent update of Forwarding Interest Base.Fog computing as an extension of cloud computing technology,has inherited from the cloud computing the flexible use and dynamic allocation of resources.Furthermore,it moves various types of computing and storage resources to the edge of the network.For applications that require real-time control and response,fog computing is a better choice because the fog node can process the data at the edge and return the results directly without uploading to the cloud.Fog computing with its abundant storage and computation resources can effectively alleviate the problem of insufficient storage capacity of network devices in the NDN and has the ability to dynamically allocate these resources as needed.Edge feature allows content caching to occur only at the edge of the network,reducing data redundancy and making more efficient use of resources.Based on the above requirements and analysis,this paper proposed a fog computing based NDN architecture,which mainly includes the following aspects: 1)the data taxonomy mechanism of NDN based on fog computing is proposed and designed.This mechanism take the use of the content-aware feature of fog nodes to categorize data into cacheable and non-cacheable data.2)Proposed and designed the mobile node management mechanism of the NDN based on the fog computing,and used the fog node as the publisher of the data to reduce the impact of the publisher's movement on the network.3)Proposed and designed the network caching optimization based on fog computing,and provided the dynamic storage resources for the intermediate node of the network by using the fog virtualization and dynamic allocation of resources.4)The control optimization of content center network based on fog computing is proposed and designed.Combined with the idea of network function virtualization and software definition network,the distributed control in the node is integrated into the fog node,and the fog node is used to control the network routing,caching and etc.