基于正交频分复用无源光网络的动态带宽分配算法研究

介绍了一种应用于未来大容量、多业务接入场景下的称为光正交频分复用无源光网络的下一代接入网技术,并提出了一种基于此结构的称为基于服务质量需求的固定周期流水线轮询动态带宽分配算法.该算法通过应用二级带宽分配机制、带宽预申请机制和最小带宽申请优先原则等方法,有效地保障了在大容量、多业务接入场景下各个业务不同的服务质量需求.为研究其性能,建立了该算法的仿真模型并进行了对比仿真.仿真结果显示该算法以低时延和低抖动有效地支持快速转发业务,同时平衡了确保转发业务和尽力而为业务之间的优先性和相对公平性,并且实现了不同光网络单元间的同优先级业务带宽分配的公平性.与传统的动态带宽分配算法算法相比,基于服务质量需求的固定周期流水线轮询动态带宽分配算法具有执行效率高、算法复杂度低、性能良好等优点,并且能够更好地适应光正交频分复用无源光网络的特性,发挥网络最大的性能.

Dynamic Bandwidth Allocation Algorithm for Orthogonal Frequency Division Multiplexing Access-passive Optical Network

An architecture of next generation access network namely orthogonal frequency division multiplexing passive optical network was presented,which has high spectrum efficiency,fine bandwidth granularity and low cost.A novel efficient dynamic bandwidth allocation algorithm based on orthogonal frequency division multiplexing passive optical network namely fixed cyclic polling in pipeline dynamic bandwidth allocation algorithm was investigated.To meet the quality of service requirements under multi-services access scenario,two level bandwidth allocation mechanisms,bandwidth pre-request for expedited forward service and the smallest bandwidth request allocated the first principle for assured forwarding and the best effort services were applied in this algorithm.A simulation was conducted to study the performance of fixed cyclic polling in pipeline dynamic bandwidth allocation algorithm in orthogonal frequency division multiplexing passive optical network.The results show that the proposed algorithm supports expedited forwarding services with low packet average delay and low jitter,balances priority and fairness between assured forwarding services and best effort services,and achieves bandwidth allocation fairness among different optical network units in the same priority.Compared with the traditional dynamic bandwidth allocation algorithms,fixed cyclic polling in pipeline algorithm meets the quality of service requirements excellently,and adapts orthogonal frequency division multiplexing passive optical network better with higher bandwidth efficiency and lower algorithm complexity.