| Network virtualization technology is one of the significant technologies proposed to solve the rigid problem that the current Internet faces in network architecture.As the key step of network virtualization technology,virtual network embedding can allocate physical resources of substrate network for virtual network requests to meet their requirements.There are two problems in the existing virtual network embedding algorithms.One is the lack of one-to-many node embedding algorithms,while the one-to-many embedding can lead to the improvement of resources utilization.The other is the lack of virtual network embedding algorithms in dynamic environment.To solve the first problem,we select the virtual private cloud embedding in data centers as the research scenario when considering the design and implementation of virtual network embedding in the static coud.In this thesis,we propose an algorithm,called greedy virtual SDN switch embedding with one-to-many embedding,together with the virtual SDN switch migration algorithm,to solve the resources allocation problem and maximizing the profit of the data center with consideration of virtual private cloud embedding in data centers.In view of the lack of virtual network embedding algorithms in dynamic enviroment,we select the vehicular cloud dynamic embedding in vehicular ad-hoc networks as the research scenario when considering the design and implementation of virtual network embedding technology in dynamic cloud.We propose the vehicular cloud embedding in vehicular ad-hoc networks algorithm and virtual node remapping to realize the dynamic embedding of vehicular cloud.This algorithm is also aimed at solving the resources allocation problem and minimizing the cost when considering vehicular cloud dynamic embedding in vehicular ad-hoc networks.Simulation results show that the proposed algorithms achieve better performance compared with existing works. |
