Research On Virtual-Network-Function Deployment For Latency Optimization In Multi-Access Edge Computing

In recent years,Multi-Access Edge Computing(MEC)has rapidly gained a lot of attention as a novel computing model in the current wave of information technology development.By extending cloud computing services to the edge of mobile networks using cloudlet(servers or server clusters)co-located with access points near user devices,which provides users with short latency and high performance computing services.In addition,Network Function Virtualization(NFV)is another promising technology.The core idea of NFV is to decouple dedicated physical devices and network functions,and instead use Virtual Network Function(VNF)implemented by flexible software technologies to replace dedicated physical devices.Virtualized network services provided in MEC can improve user service experience,simplify network service deployment and streamline network resource management,moreover,instantiating VNF at the network edge can help reduce service latency and improve bandwidth utilization.Although the combination of MEC and NFV technologies efficiently solves the structural problems caused by the diversity of edge devices,the challenges will be faced with how to deploy VNFs to meet the high performance computing requirements of network tasks while also reducing the task service latency and how to allocate limited resources to each application to achieve high resource utilization.To this end,based on the above challenges,the main work of this paper includes as follows:1)To address the challenge of how to deploy VNFs to meet the high performance computing requirements of network tasks while considering computing resources and reducing task service latency,A MEC scenario is constructed,network,task,communication,and computation models are built,and Mixed-integer Nonlinear Programming(MINLP)problem on VNF deployment and resource allocation is proposed with the aim to minimizing latency and consideration of computing resources.And we demonstrate that is NP-hard problem.2)In order to solve the above problem effectively,a Set Covering Problem(SCP)-based deployment scheme is designed to solve it.It is solved by converting the original problem into an SCP.A time complexity analysis is also performed.3)To solve the optimization problem effectively,this paper further considers the optimization of computational resources and splits the model into two subproblems,namely,the VNF deployment subproblem and the computational resource allocation subproblem.It is also proved that the computational resource optimization problem is a convex optimization problem,and the optimal solution is found using the Lagrange multiplier method.Then an approximation algorithm based on the original dual is designed to solve the problem,and an algorithm analysis and proof are performed,which includes the proof of the approximation degree and the time complexity analysis of the proposed algorithm.4)The effectiveness of the proposed algorithm in this paper is verified by comparing the delay values and running times of the five algorithms under different parameters,highlighting the advantages of the proposed algorithm in terms of low latency and high resource utilization.