Research On Multi Demensional Optical Network Virtualization

With the growing demand of new network services, the characteristics of large capacity and flexibility issues are more and more important. On the one hand, the current fiber resources have been exhausted in the frequency domain because of the large capacity characteristic, the transmission capacity of the single core fiber has been approaching the Shannon limit, so there is an urgent need to further improve the transmission capacity of the optical fiber communication system from the space domain. On the other hand, the advance reservation request with obvious time and flexible character is appearing. To conquer these challenges, optical network virtualization technology can slice the substrate resources to meet the different requirements, by abstracting and cooperatively allocating the underlying time, space and frequency resources,allowing customers to use the underlying physical resources as a service without realizing the implementation details. It can improve the multi-dimensional resource utilization.Based on the large capacity and the flexibility characteristics, we propose multi-dimensional optical network virtualization. This thesis studies the resource allocation spatial division multiplexing elastic optical networks, and the resource collaboration of time and spectrum domain is studied. The main works of this thesis are as follows:Firstly, aiming at the problem of multi-dimensional resource co-optimization of frequency domain and space domain in space-division multiplexing optical network, a spectral resource description model based on crosstalk is established based on the characteristics of space-division multiplexing optical network, crosstalk-aware spectrum compactness (CASC).Based on this resource description model, two types of routing, core and spectrum resource allocation algorithms (RCSA) are proposed, the first-fit crosstalk-aware spectrum compactnesss RCSA algorithm (FF-CASC) and the random fit crosstalk-aware spectrum compactness RCSA algorithm (RF-CASC). The numerical simulation results show that FF-CASC and RF-CASC can achieve lower blocking probability and higher resource utilization compared with the traditional crosstalk aware RCSA algorithm. FF-CASC can decrease the blocking probability 39%.Secondly, according to the concept of virtual optical network mapping in space-division multiplexing optical network, this paper proposes a method based on network element mapping in space-division multiplexing optical network by introducing the concept of network element mapping capability,namely link embedding capability and node embedding capability. Based on this concept, two virtual optical network mapping methods are proposed, the adjacent link embedding algorithm (ALEA) and the remote link embedding algorithm (RLEA). In the simulation part,this two mapping algorithms are compared with the basic virtual optical network mapping algorithm (Baseline).The simulation results show that compared with Baseline, the proposed virtual optical network mapping algorithms based on network element mapping capability can achieve lower blocking probability and higher cost-benefit ratio.RLEA can decrease the blocking probability 40% and increase cost-benefit ratio 8%.Thirdly, according to the virtual optical network mapping problem of advance reservationrequests in elastic optical network, this thesis first designs the two-dimensional resource measurement method based on the characteristics of advance reservationrequests, exploring the time and spectrum two-dimensional resource fragmentation. Based on this method, to reduce the two-dimensional resource fragmentations and to solve the problem of insufficient resources, this thesis proposes a multi-path fragmentation-aware routing,modulation format, and spectrum allocation algorithm. Based on the two-dimensional resource compactness measurement method, this thesis proposes a two-dimensional fragmentation aware virtual optical network mapping algorithm (TDFA) and fragmentation-aware and continuity-aware virtual optical network mapping algorithm (FCA). The numerical simulation results show that TDFA and FCA algorithm can effectively reduce the traffic blocking probability and improve the utilization of the network resources in different durations of the services or different simulation topologies.