为了减少弹性光网络(EON)中的频谱和时间碎片, 针对提前预留(AR)业务提出一种碎片感知路由与频谱分配(RSA)算法。在选路阶段, 该算法采用K最短路径算法, 综合考虑路径距离、跳数、频谱资源消耗和邻接链路数等因素影响来挑选侯选路径集；在资源分配阶段, 则采用局部影响因子统计可能产生的频谱和时间碎片, 以降低候选资源块周围的频谱、时间资源碎片化程度, 并利用整体影响因子尽可能将占用资源集中在频率轴边界处, 以增加中间空闲资源连续性, 同时还引入时域资源使用度因子以避免过度占用资源。仿真结果表明, 该算法进一步改善了带宽阻塞率性能, 提高了资源利用率。

In order to reduce spectrum and time fragmentation in elastic optical networks (EON), a fragmentation-aware routing and spectrum allocation (RSA) algorithm is proposed for advance reservation (AR) services. In the stage of route selection, the algorithm uses K-shortest path algorithm to select alternative path set by comprehensively considering the influence factors of path distance, hop count, spectrum resource consumption and adjacency link number. In the stage of resource allocation, local influence factors are used to count the possible spectrum and time fragmentation, so as to reduce the fragmentation of spectrum and time resources around the candidate resource blocks. The overall influence factors are used to concentrate the occupied resources at the boundary of the frequency axis as much as possible, so as to increase the continuity of intermediate idle resources. At the same time, time-domain resource utilization factors are introduced to avoid excessive resource occupation. The simulation results show that the algorithm improves bandwidth blocking probability performance and resource utilization.