Research On Resource Optimization Of Latency-Sensitive Services In Collaborative Cloud-Edge Elastic Optical Networks

With the rapid development of Internet of Things(IoT),smart city,face recognition,intelligent transportation system and other emerging applications spring up.Traditional cloud computing and edge computing can no longer meet the requirements of high speed,low latency,and high reliability of emerging applications.Therefore,in order to better meet the requirements of applications,the concept of collaborative cloud-edge elastic optical networks(CCE-EONs)is proposed to provide the computing resources and bandwidth resources,which is co-processed by cloud computing and edge computing.This paper focuses on the end-to-end(E2E)latency of computation offloading,minimizes the overall latency of latency-sensitive services as much as possible,and optimizes the allocation of network resources in CCE-EONs.Firstly,in order to optimize the deployment of the edge computing servers in CCEEONs,a hierarchical deployment network framework is proposed,and then a hierarchical deployment algorithm is developed based on path importance degree for resource optimization in CCE-EONs.The simulation results show that compared with traditional offloading algorithms,the proposed hierarchical deployment algorithm can effectively reduce E2E latency for resource optimization in CCE-EONs.Secondly,for the latency-sensitive service resource offloading strategy in CCE-EONs,two latency-sensitive network resource offloading integer linear programming(ILP)models are established for network resource offloading,and then a network resource offloading algorithm based on both latency and resource-equilibrium is proposed to reduce latency and balance traffic load.Simulation results show that compared with the existing algorithms,the network performance of the proposed algorithm is closest to that of ILP models,which reduces E2E latency and optimizes the allocation of network resources.Finally,aiming at solving the problem of offloading strategy of different service types in CCE-EONs,on the one hand,a partial resource offloading ILP model for service classification is established.On the other hand,according to different service types,the concepts of offloading decision variables and server importance degree are introduced,and then the optimal proportion segmentation is deduced.Based on this,a partial resource offloading heuristic algorithm based on proportion segmentation is proposed to improve the resource efficiency.Simulation results show that compared with the traditional partial resource offloading algorithm,the results of the proposed partial resource offloading algorithm are closest to the ILP model,and E2E latency can be further reduced in CCEEONs.