Research On Virtual Network Embedding Method

As an important technology to overcome the rigidity of network architecture,network virtualization has attracted more and more attention and research.By sharing infrastructure,network virtualization allows the deployment of multiple heterogeneous networks in physical networks to meet the diverse needs of people for network services.In addition,because the infrastructure is used together by multiple virtual networks,the utilization rate of physical network resources is greatly improved,and the inefficient use of infrastructure is avoided.A major challenge in implementing network virtualization is how to allocate physical resources to virtual networks effectively,that is,to embed virtual networks in physical networks.This is called the problem of virtual network embedding.At present,research on virtual network embedding methods is also increasingly focused..The problems that need to be solved in virtual network embedding are divided into two aspects: one is the embedding of virtual nodes,and the other is the embedding of virtual links.The traditional embedding algorithm mainly embeds the virtual nodes according to the node CPU resources.However,this ignores the link resources around the nodes,which affects the success rate of virtual link embedding.In addition,when the virtual link embedding is performed,since the general physical link weight is set to a fixed value,part of the link resources are quickly exhausted,and thus the virtual network embedding will be initially rejected.For this reason,this paper mainly studies the virtual network embedding algorithm from the aspects of virtual network acceptance rate and resource usage.In the cable network,this paper analyzes and summarizes the current virtual network embedding algorithm,proposes an algorithm GR-SP-SAA based on node resources embedded in the virtual network,and then uses the simulated annealing algorithm to optimize the use of links,thereby effectively improving the acceptance of virtual network requests.rate.In addition,by defining link weights,it is ensured that the physical links with more available bandwidth resources will be preferentially selected at each embedding,thus realizing the dynamic adaptability of the algorithm.The simulation results show that the GR-SP-SAA algorithm has a higher acceptance rate andis more efficient in the use of physical resources.There are additional challenges for virtual network embedding due to the presence of link interference in the wireless network.This paper proposes a coordinated MWL-SP algorithm and embeds virtual nodes and virtual links at the same time,effectively avoiding the interference area.In addition,the link weight is redefined in combination with the characteristics of the wireless network,reflecting the dynamic adaptability of the algorithm.Simulation results show that the MWL-SP algorithm is extremely efficient for the use of physical resources and significantly improves the benefits of the physical network.