Abstract: Under the constraints of spectrum consistency and spectrum continuity, after a period of operation of elastic optical network, a large number of fragments of network spectrum will appear. To address the problem of fragmentation and unbalanced allocation of service among nodes, a routing and spectrum allocation algorithm based on node importance is proposed. Aiming at routing problem, this algorithm considers the type and size of service, finds the key nodes in the network, and makes service allocation balanced. In terms of spectrum allocation, the algorithm can reduce the spectrum fragmentation as much as possible by taking into account the distribution of spectrum resources on the network link and combining the number of frequency slots required for each service. In two network topologies of different sizes, NSFNET and USNET, the proposed algorithm is simulated. The simulation results show that the proposed NIRSA algorithm can not only effectively reduce the traffic blocking rate, but also improve the spectrum utilization rate of the network and realize the improvement of network performance.

Key words: elastic optical network, load balancing, spectrum resource utilization rate, blocking rate

摘要： 在频谱一致性和频谱连续性的约束条件下，弹性光网络运行一段时间后，网络频谱会出现大量碎片的问题。针对碎片以及业务在各节点间分配不均衡的问题，提出了一种基于节点重要度的路由选择与频谱分配算法NIRSA（Route Selection and Spectrum Allocation algorithm based on Node Importance）。该算法针对路由选择问题，考虑业务的类型与大小，找出网络中的关键节点，使得业务分配达到均衡。在频谱分配方面，算法考虑到网络链路上频谱资源的分布情况，结合每个业务所需的频隙数，可以尽可能地减少频谱碎片。在NSFNET和USNET两个不同规模的网络拓扑环境下，对所提算法进行了仿真实验。仿真结果显示，所提出的NIRSA算法既可以有效地降低业务阻塞率，又能提高网络的频谱利用率，实现网络性能提升。

关键词: 弹性光网络, 负载均衡, 频谱资源利用率, 阻塞率