Research On Routing Strategy In Icn-based Computing First Network

With the rapid development of the Internet’s high-tech,emerging businesses such as digital twins,long-range medical care,metaverse,and Web3 have gradually emerged,which have put high requirements on computing power resources and network transmission efficiency.However,the traditional cloud computing architecture and edge computing architecture cannot meet their needs.Therefore,the computing power network is proposed as a new type of network architecture.However,there are some problems with the construction of computing power network architecture.For example,the computing power resources of computing power nodes are heterogeneous,computing power services are diverse,user task needs are unique,and network link loads will change dynamically.Therefore,this thesis introduces the information center network(ICN)to build an ICN-based computing first network architecture to realize the realtime perception of the network resources,the unified scheduling of user tasks,and the routing of data transmission.The specific work of the paper is as follows:1)This thesis proposes a multi-task routing and scheduling scheme based on resource-awareness of computing network.First,we propose an ICN-based computing power network architecture,designs a hierarchical function module,and illustrates the role of various levels of entity.In addition,this thesis integrates ICN’s layered naming mechanism,routing mechanism and cache mechanism,and computing power network,helping to achieve computing power perception,service request route forwarding,and intranet service container cache.Based on this network architecture,this thesis will be considered in cooperative considerations such as differentiated user needs,heterogeneous computing power and load status,path planning,and task scheduling order,and will be integrated by control nodes to reduce the average service delay and system service energy consumption of the user.In response to this optimization goal,this thesis proposes a solution based on the Dueling Double Deep Q-Network(D3QN)algorithm and compares it through simulation experiments with common solutions.The results show that the algorithm has a better performance.2)This thesis proposes a delay-optimized routing strategy based on dynamic service caching.First,we propose an intra-network service container cache model.The node will replace the service docker according to the default cache strategy.Under this model,this thesis considers different link network bandwidth and dynamic load changes,select the optimal path for user tasks,forward the task to the network computing node or edge calculation node,and solves the delay of the minimum user service.In response to the optimization goal,this thesis proposes a deep reinforcement learning algorithm based on deep definitive strategy gradient(DDPG)and compares multiple dimensions such as the common scheme calculating nodes,node performance,and routing nodes from the common solution.The dimension compares.Experimental results show that the algorithm proposed in this article has a better performance.