Research On Latency-Aware Virtual Network Embedding In Multi-Dimensional Optical Network

Faced with the current explosive increase in the number of network users and transmission devices,how to meet the corresponding massive network transmission bandwidth requirements is a serious problem.With the advent of space division multiplexing technology,optical fiber resources are no longer limited to frequency and time domain dimensions.Multi-dimensional optical networks relying on space division multiplexing technology and elastic grid technology alleviate the problems of insufficient network capacity.However,multi-dimensional optical networks The lack of network resources has also resulted in the rigidity and complexity of resource management in optical networks.In this case,efficient and flexible management of multi-dimensional optical networks can be achieved through optical network virtualization technology,in which the network virtual topology mapping problem can provide end users with higher-quality services on the basis of improving the utilization of physical network resources..According to the Cisco Visual Networking Index report,there are more and more scenarios that require high immediacy.Therefore,this paper studies the virtual topology mapping problem of multi-dimensional optical network delay awareness,and combines the multi-dimensional optical network environment and delay characteristics to build a virtual topology mapping problem model.Link distance adaptation scheme and transmission delay calculation model,and then two delay-aware virtual topology mapping algorithms are proposed.In Chapter 3,a multi-attribute constraint cross-virtual topology mapping algorithm MCC-VNE is proposed.Different from the traditional mapping algorithm that only considers the computing resources and bandwidth resources of node links,the priority of services,nodes and links in the mapping process is comprehensively considered from multiple attribute dimensions,and the corresponding formula is designed.In addition,the MCC-VNE algorithm is different from the two-stage sequential mapping of node links in the past,which cross-maps links and nodes,and the intermediate mapping results will affect subsequent mappings.The simulation results show that,compared with the traditional algorithm,the MCC-VNE algorithm not only improves the virtual network request acceptance rate,but also significantly reduces the average virtual network delay,and at the same time ensures load balance and saves resource costs.In Chapter 4,this paper proposes a candidate subset and delay-aware virtual topology collaborative mapping algorithm CSLA-VNE based on integer linear programming model.On the basis of the mathematical models in the previous two chapters,the heuristic idea is adopted to reduce the time complexity of the algorithm by first establishing a subset of node link candidate mappings.After that,the CSLA-VNE algorithm is based on the integer linear programming ILP model,and the objective function and constraints are constructed to find the optimal solution with the minimum average delay of the virtual network.The simulation results show that,compared with the traditional heuristic and precise MIP algorithm,the CSLA-VNE algorithm significantly reduces the average virtual network delay,and also improves the request acceptance rate and resource utilization,which verifies the effectiveness of the CSLA-VNE algorithm.