The Shortest Delay Routing Algorithm With Computational Task Constraints

In recent years,as new business needs continue to be proposed,new networks have also sprung up,and they have gradually become diversified and personalized.Currently,the more common new networks include the Internet of Things,the Internet of Vehicles,and smart cities.These new services require the network to have computing capabilities,which has promoted cloud computing and edge computing to gradually become research hotspots in recent years,and further promoted computing resources to become the basic resources of the network.This paper mainly focuses on the communication network with computing capability,and studies the shortest delay routing algorithm with computing task constraints.Computing tasks can be divided into two types.The first is a computing task that cannot be divided.The data must be concentrated on a node to perform computing services as a whole,such as image imaging or recognition;the second is a computing task that can be divided.Distribute the data on multiple nodes and perform computing services at the same time,such as searching for sensitive information in the data.Therefore,for these two types of computing tasks,this paper proposes an end-to-end shortest-latency routing algorithm for single-node computing services and an end-to-end collaborative computing service for multiple nodes,combining the processing delay of nodes and the complete arrival delay of end-to-end data.The shortest delay routing algorithm to the end.The main work of this paper can be described as follows:First,introduced two common shortest delay routing algorithms,and then analyzed the network topology model used by these two algorithms,and found that the model cannot meet the shortest delay routing algorithm that introduces computational tasks and capacity constraints.Research.Therefore,for data transmission services with computing requirements in communication network scenarios,the maximum transmission rate of network links and the representation of node computing capabilities are added to the original network topology model,thereby constructing a network topology model for joint computing.At the same time,the demand constraint of the computing task is added to the transmission demand of the task,thereby constructing the task characterization model.The newly constructed network model and task model will pave the way for subsequent research on the shortest delay routing algorithm with computational task constraints.Then,for the indivisible computing tasks,an end-to-end routing algorithm with the shortest delay is proposed in which the computing tasks are concentrated on one node for computing services.The first step is to calculate the complete arrival delay of a given amount of data under a given path in the network topology model of joint calculation without considering the calculation scenario.The given path here may be either a single path or There may be multiple paths.The second step is to study the single-path shortest complete arrival delay routing algorithm and the multi-path shortest complete arrival delay routing algorithm for a given data volume in this scenario.The third step is to take calculation into consideration,traverse all nodes with computing capabilities in the network,and find the computing node with the shortest delay required to complete a given task.Combining the shortest route from the source node to the computing node and the shortest route from the computing node to the destination node is the end-to-end shortest delay route where the required computing tasks are concentrated on one node for computing services.Finally,through MATLAB simulation,the effectiveness of the algorithm is verified.Finally,for computing tasks that can be divided,an end-to-end routing algorithm with the shortest delay is proposed in which computing tasks are distributed among multiple nodes for collaborative computing services.Since the computing task can be executed on all nodes with computing capabilities that the data transmission path passes through,while transmitting data,all nodes with computing capabilities on the path are allowed to perform calculations within the time range of their effective execution of computing services.In this way,the shortest time delay to complete a given task under a given path is obtained.Then it traverses all possible paths and finds the end-to-end shortest delay route in which computing tasks are distributed across multiple nodes for collaborative computing services.Finally,the effectiveness of the algorithm is verified by MATLAB simulation.