Research On Task Offloading For Cost Optimization In Multi-access Edge Computing

Multi-Access Edge Computing(MEC)is a new network architecture that provides powerful computing capacity and large amounts of storage space for users.It can push computational resources closer to users and enable compute-intensive and low-latency tasks to be handle efficiently.In response to the heterogeneous nature of edge-cloud networks,Network Functions Virtualization(NFV)helps reduce the difficulty of deploying edge network environments.It enables Virtual Network Functions(VNFs)to run in the software by separating network functions and services from dedicated network hardware.The combination of MEC technology and NFV technology optimizes the architecture of the edge-cloud network.It can provide real-time intelligent processing and execution to users' local tasks.However,the cost consumption of edge-cloud networks is rarely considered in current research,which is particularly important for operators providing network services to reduce cost consumption and thus ensure revenue in order to achieve technological innovation and sustainable development.Therefore,in order to reduce the cost consumption of edge-cloud networks,the main work of this paper includes as follows:(1)To address the problem of how to use and deploy VNF instances to meet the users' requirements of Quality-of-Service,the models of the MEC network,tasks,communication and cost consumption are built.Aiming to minimize the cost consumption of the edge-cloud network,a cost optimization problem under computing resources and time latency constraints is proposed by proved to be a NP-hard problem.(2)To obtain an effective solution to the above problem,considering that the NP-hard problem is difficult to find an optimal solution,a heuristic algorithm is proposed to transform the original problem into a bin packing problem,which reduces the cost consumption of the edge-cloud network to a certain extent.(3)In order to solve the optimization problem,this paper integrates the cost consumption of transmission,VNF instantiation and task execution,and proposes an approximation algorithm based on the heuristic algorithm with the idea of dynamic programming to reasonably allocate the computing resources in the edge-cloud network and minimize the cost consumption of the edge-cloud network.And the approximation ratio is proved.(4)Finally,simulation experiments are conducted.The two algorithms proposed in the paper are compared with random algorithm and greedy algorithm with each other.The heuristic algorithm has the shortest runtime,the approximation algorithm can always offload the most tasks by using limited computing resources and reduce the cost consumption to the greatest extent.