Font Size: a A A

Research On Resources Management Based On Network Virtualization

Posted on:2018-03-09Degree:MasterType:Thesis
Country:ChinaCandidate:T T ZhaoFull Text:PDF
GTID:2348330521450995Subject:Engineering
Abstract/Summary:PDF Full Text Request
In high information modern society,people rely on the network is also r eflected in all aspects of life,Therefore,many newly arisen Internet services have emerged.Traditional Internet network frame can't fulfill development of the modern network.In other words,an urgent problem that the Internet rigidity came out.At this point,the network virtualization technology is proposed.It allows several logically networks can coexist in a physical network using measures such as abstraction and isolation.Each logical network can its own topology and they are isolated from each o ther.So Each logical network can run its own protocols,deploy its own business,serve different users.Therefore,network virtualization technology is an effective method to solve the shortcomings of the traditional network at present.At present,the hot issues of network virtualization research are virtual network mapping problem and resource allocation problem.Virtual network mapping can be described as a process for requesting a virtual network to find nodes and links that meet the resource requirements in the underlying physical network.Selecting reasonable physical nodes and links for virtual network nodes and links will map more virtual network requests to physical networks and save the mapping time.In network virtualization,multiple virtual networks coexist on a physical network.But physical network resource is finite,the effective resource al ocation algorithm is needed.In this paper,the dynamic allocation algorithm is divided into two stages.The first stage is the initial stage of the resource,that is,the virtual network mapping phase,and the second stage is the stage of dynamically adjusting the allocated resources after the virtual network mapping ends.During the virtual network mapping phase,analysis the classic virtual network mapping algorithm and point out the advantage and weaknesses firstly.Aiming at these disadvantages,virtual network mapping algorithm which evaluate the characteristic of the node comprehensively is proposed.The algorithm selects the nodes of the virtual network and the nodes of the physical network with different evaluation indexes.During the linking mapping stage,because the effect of the resource on the link connected to the node hasbeen taken into account when the node is selected,therefore,linking mapping will do the mapping according to the K shortest-path algorithm and it will ease the mapping process and reduce the mapping time.In the simulation part,several algorithms are compared and analyzed in the virtual network request acceptance rate,virtual network revenue,cost and virtual network revenue and cost ratio.Through the comparative analysis,it can be seen that the virtual network mapping algorithm studied in this paper has a certain increase in cost while improving the profit,but the whole has good optimization effect.In the study of dynamic resource allocation algorithm,we set different weight coefficients for the virtual network with different priorities according to the the idea of weight coefficients vector and multi-objective programming,and so that the virtual network with higher priority will hava more opportunities to the network resources.When the requested virtual network resources change,it can adjust the weight coefficient timely and ensure the resources can be adjusted between the different virtual networks.In the simulation,the dynamic resource allocation algorithm has better resource utilization for the physical network by analyzing and comparing the link bandwidth utilization and the virtual network satisfaction of the static resource al ocation and the dynamic resource al ocation algorithm.
Keywords/Search Tags:network virtualization, virtual network mapping, node mapping, link mapping, dynamic resource al ocation algorithm
PDF Full Text Request
Related items