Research On Dynamic Network Slicing Optimization In Metro Optical Network

Today,with the rapid development of network technology,new applications are continuously emerging,enriching and facilitating our daily lives.Meanwhile,the vision of new applications for the future is constantly raising new requirements for the underlying network infrastructure.As a bond between the access network and the core network,the metropolitan area network(MAN)carries more and more applications and deploys more and more computing infrastructure.It is the core of future network upgrades.For upper-layer users to use underlying metro network resources more conveniently,network slicing that integrates multi-layer network and computing resources is an important solution.For the user,the slice service should have high configuration flexibility,reliability,and controllable service interruption.For network operators,the provisioning of slicing services should consider resource utilization and resource allocation efficiency,so as to achieve higher system operation efficiency and enhance the profits of service supply.Therefore,considering the above concerns for slicing services,this research focuses on the topic of dynamic MAN slicing and proposes a set of staged optimization schemes for heterogeneous multi-level MAN slicing.First of all,this paper summarizes the new trends in the current network development,reviews the research progress in the field of optical network resource allocation,virtual resource embedding and some other related topics.Then,the research significance of topics including survivability in dynamic network slice resource allocation,joint optimization with computing resources,resource reconfiguration,and service migration based on elastic filterless optical network is introduced in detail,as well as the research basis of this research,such as network architecture,optical network resource characteristics,and supporting technologies for slices.Next,for the creation of dynamic slices,this research focuses on the establishment of fiber trees before resource allocation in filterless optical networks and the embedding of virtual resources during the construction of actual slice resources,and thus constructs heuristic solutions based on genetic algorithms for both.For the former,the optimizing pertinence and stability of demand-aware and balancing strategy are mainly verified.For the latter,survivable elastic sharing,and computing resource balancing optimization are innovatively proposed,thereby improving the resource allocation efficiency,flexibility,and reliability of the system.In addition,the runtime reconfiguration of dynamic slices is considered.For one thing,for the collaborative migration of computing and communication resources,which has not been discussed in detail in the research area,this research proposes a migration bottleneck-aware dynamic service sequential migration strategy based on the creation of multi-path parallel migration channels,greedy algorithm ideas,and artificial bee colony algorithm,which improves the migration efficiency of computing resources within a slice while guaranteeing uninterrupted operation of upper-layer services.For another,a resource migration-aware network slice remapping scheme is proposed to meet the adjustment of slice requirements while reducing the misalignment between old and new slices to avoid unnecessary resource reconfiguration.Simulation results verify the resource optimization of the scheme mentioned above.