Font Size: a A A

Research On Network Virtualization Method In Multi-tenant SDN Data Center Network

Posted on:2017-05-13Degree:MasterType:Thesis
Country:ChinaCandidate:Y M ChenFull Text:PDF
GTID:2348330518495867Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
The development of cloud computing and data center network multi-tenant demand,put forward higher request for the data center network.Traditional data center network is lack of unified management and distribution of network resources,it is difficult to meet the network requirements of different tenants.Therefore,the industry and researchers have proposed pooling the physical network resource,using virtualization technology,to meet the demands of special networks of different tenants.However,the existing data center resources allocation is mainly based on the virtual machine.The lack of effective management of network resources,which is difficult to ensure that the network requirements of different tenants and the underlying physical network resource efficient use.To solve this problem,some of the previous studies focus on using heuristic algorithm to solve the problem of resource allocation,which can effectively solve some problems,such as the energy consumption problem of the data center network resources,load balancing problem,multipath problem,and so on.But these algorithms are mostly in traditional network environment,and without considering the differences of nodes and links.Software Defined Network(SDN)network has the characteristics of centralized control,network virtualization,open interface,and so on,which can well fit the demands of network centralized management and flexible networking in multi-tenant data center network.SDN network logical virtualizes the physical network and centralized manages them,which provides a new way for resource allocation.Given the characteristics of the physical network resources in SDN network are different from traditional network,the existing researchers have put forward some virtual network embedding algorithms in SDN network environment,which solve some problems,such as load balancing problem in SDN network,dynamic adjustment of resource allocation,and so on.But in these algorithm,the key nodes or key link in physccial network are easily to be taken up by the first virtual network requests and become bottlenecks,thus cause the deterioration of the physical network resources utilization.Therefore,it is necessary to further research on the virtual network embedding algorithm in SDN network environment.Put forward a virtual network mapping algorithm to avoid the key nodes or key link become bottlenecks in the virtual network mapping,thus effectively improve the utilization rate of the physical network resources.In this paper,we proposed a topology-aware virtual network embedding algorithm in the SDN network environment.Characteristics of the substrate network resources in the SDN network environment and differences of physical nodes and links in the substrate network are introduced into the virtual network embedding algorithm.We defines the virtual network embedding problem into a linear programming,which ensures the time complexity of the algorithm is polynomial time.The simulation experimental results show that the proposed virtual network embedding algorithm effectively improves the virtual network request acceptance ratio,the average revenue,the average utilization of the substrate network resources and the revenue/cost ratio,and ensure the average cost is low at the same time.Firstly,this paper analyzes the current research results and the existing problems of the virtual network embedding in SDN network environment;Then,introduce the prosed topology-aware virtual network embedding algorithm in the SDN network environment in detail;Finally,this paper introduces the design and realization of the simulation platform,and evaluate the performance of the proposed algorithm.
Keywords/Search Tags:multi-tenant data center network, network virtualization, virtual network embedding, SDN
PDF Full Text Request
Related items