Research On Virtual Optical Network Oriented Spectrum Trading Scheme

With the rapid growth of new emerging Internet services such as virtual reality and high-definition video,as a backbone network,the elastic optical network is facing tremendous pressure.To improve the traffic-carrying capability of the EON,network virtualization is applied.Through dividing a physical optical network into multiple independent virtual optical networks(VONs),network virtualization can effectively improve the flexibility of service provisioning and the utilization of network resources.However,in most VON resource allocation approaches,resources allocated to each VON are often fixed during its lifetime,while its carried traffic demand is always dynamic.This would lead to the mismatch between the allocated spectrum resources and the carried traffic demands.When the traffic demand on a VON is low,some of its allocated spectrum resources remain idle and wasted.On the contrary,when the carried traffic demand on a VON is high,its allocated spectrum resources are insufficient.To address the above mismatch,we propose a new mechanism called spectrum trading(ST)that enables VON clients to form a cooperative community and allows them to trade spectra with each other according to their actual bandwidth requirements.Specifically,VON clients with unused spectra can trade them to other VON clients with excessive traffic demands so that the unused spectra of the former can be used and the excessive traffic of the latter can be carried.Based on the spectrum-trading problem,we develop two mixed integer linear programming(MILP)models and two efficient heuristic algorithms.Simulation results show that the proposed ST scheme can significantly increase total carried traffic demands on VONs and the overall utilization of spectrum resources in the physical network.The ST scheme can improve the total carried traffic demands on all the VONs.However,for a situation that the total allocated spectrum resources in the community are not enough to serve all traffic demands on the VONs,resource shortage of the VONs cannot be addressed through spectrum trading.To tackle this,we propose an approach to purchase extra spectrum resources from the network operator.Also,to reduce the cost of extra spectra that the VON clients need to purchase,we propose a cost-minimized spectrum trading(CMST)scheme that enables VON clients to purchase extra spectra from the network operator with minimum cost if they cannot obtain enough resources through the ST scheme.Based on this,we further develop an MILP model and corresponding heuristic algorithms.Simulation results show that the proposed CMST scheme can significantly reduce the overall cost paid by all the VONs for purchasing extra spectrum resources.