Virtual Network Function Resource Allocation And Service Function Chain Routing

The delivery of end-to-end services usually requires a variety of network functions to process traffic,which constitutes the concept of traditional service chain.Network function(such as firewall,deep packet detection,video acceleration,etc.)can filter and detect traffic,which plays an important role in the network.However,in the tra-ditional network,network functions and hardware devices are closely coupled,and the deployment and change process of service chain is very rigid,which can not meet the diverse needs of users.Network Function Virtualization(NFV)decouples hardware and software,and realizes network functions in the form of software,which is called Virtual Network Function(VNF).Virtual network function can be created dynamically and deployed on general server devices flexibly.Software Defined Networks(SDN)separates the control plane from the data forwarding plane of the network,and uses a centralized controller to automatically distribute the flow table of the service function chain.The service chain based on SDN/NFV technology is called Service Function Chain(SFC).We need to find a complete path between the service terminal and the user in the underlying physical network.This path needs to meet the physical node re-source requirements of the virtual network functions and the bandwidth requirements of the virtual links in the service function chain.This problem is called the virtual network function deployment,resource allocation and the the service function chain routing.In this article,DP-COA,a cost optimization algorithm based on dynamic program-ming,is proposed to solve the problem of virtual network function deployment and traffic forwarding.First of all,we consider the running cost of nodes,the deployment cost of virtual network functions and the communication cost jointly,and establish a binary integer programming model to minimize the total cost of the network.Then,according to the characteristics of service function chain,the problem is regarded as a multi-stage decision process,and DP-COA is used to solve the problem.Finally,the simulation results show that DP-COA can effectively reduce the total network cost,and its advantages are more significant with the increase of network resource utilization.The acceptance rate of service function chain request of DP-COA is 10%-20%higher than that of other algorithms,while ensuring high network throughput.Aiming at the problem of resource allocation of virtual network function and rout-ing of service function chain,we propose a heuristic algorithm MRAA based on mini-mum resource allocation.Considering the the relationship of between virtual network function resource demand and network load,a convex optimization model is established based on the concept of queuing network.The optimization goal of the model is to find the optimal resource allocation scheme for the deployed virtual network function in-stances in the network.Then,we use MRAA to solve the resource allocation scheme,and use an open-source discrete event simulation tool Ciw to simulate the queuing,and the physical resource consumption,average packet delay and packet delivery rate are evaluated.The simulation results show that in terms of resource consumption,MRAA consumes 3%-10%more resources than the exact solution of the optimization model,but it has great advantages over the conventional resource allocation scheme.At the same time,the packet delivery rate performance of MRAA reduces by less than 1%,and the end-to-end delay of packets in the service chain is also guaranteed.