一种自适应光突发装配映射机制及性能仿真

An Adaptive Map Algorithm to Support QoS for Optical Burst Assembly

分组粗颗粒分类因有良好的扩展性,可作为OBS(Optical Burst Switching)网络的QoS(Quality of Service)实现基础。然而OBS网络因其固有限制而只能支持较少的突发优先级,因此OBS常采用多类别光突发装配技术。该文在粗颗粒分类基础上,提出了一种新的适用于这类装配的映射机制TQ-Map,即确定某一优先级光突发可以装配哪些类别分组。自适应性是该机制的突出特点,主要体现在:映射关系能随各类分组的QoS要求和负载的改变而自动进行调整。仿真表明:TQ-Map机制的自适应性能有效保证突发装配与上层QoS要求相匹配(以分组损失率为衡量指标)。

Because of its scalability, the idea of coarse packet classification can be utilized to OBS network. However, due to the limited priorities supported by OBS networks, multi-class burst assembly technique is usually adopted. Hence, a novel map algorithm, i.e. TQ-map, is proposed, which can determine which classes of packets a priority burst can contain, and fit for the multi-class burst assembly. Its most important feature is adaptivity. Compared with S-map, TQ-Map can establish the map relations varying with both traffic condition and QoS requirement, and moreover allocates the burst assembling capacity between packets' classes fairly, efficiently and differentially. The only one purpose is to match the QoS requirements of high layer's applications. The simulation results show that TQ-Map can effectively achieve the purpose to support QoS.