基于延时状态统计的分布式QoS路由算法

为了克服网络状态的不准确性给服务质量(QoS)路由机制带来的负面影响,针对延时约束最小代价(DCLC)路由问题,提出了一种基于延时状态统计的分布式路由算法(DSS-DRA)。根据延时状态的稳定性程度预测路径状态的不准确性,并结合概率选路,为到来业务探测寻找满足其QoS要求的DCLC路径。仿真及其分析表明,DSS-DRA能够较好地容忍不准确的网络状态信息,在路由性能以及DCLC问题的解决能力方面具有较为明显的优势,同时与泛洪探测等方法相比,可以较好地减小探测开销,而与最小延时等算法相比,当网络重载时,能够以略微增加的开销为代价换取较大的性能改进。

Delay state statistical distributed routing algorithm

The negative influences of the Network state uncertainty on the quality of service (QoS) routing mechanisms are reduced by a delay state statistic-based distributed routing algorithm developed in view of the delay-constrained least-cost (DCLC) routing problem. Combinal with probability routing, the algorithm analyzes the state stability to predict the network delay state uncertainty and based on which is used to probe the DCLC paths for incoming traffic. Simulations indicate that the system can tolerate imprecise state information and improves routing performance and the ability to deal with the DCLC problem. Compared with other distributed probing algorithms like flooding, the algorithm has relatively less probing overhead. Compared with algorithm based on the least delay path, it provides much better performances at the cost of a small increment in overhead when the network load is heavy.