论文部分内容阅读
OBS环境较为复杂,为了降低丢包率,减少网络阻塞和突发包冲突的概率,设计了一种OBS考虑优先级的突发包碎片检测方法,介绍了OBS基本原理和OBS考虑优先级的突发碎片传输过程,在突发包在核心节点出现冲突的情况下,依据突发包的优先级选择分段丢弃的方式,完成分片后,目的核心节点将传输反馈信息至边缘聚集节点。构建自相似模型,介绍了分形布朗运动过程的统计特性,结合自相似模型与分形布朗运动过程的统计特性对OBS中突发包碎片特征进行提取。选用Gentle Adaboost(GAB)分类算法,给出弱分类器,介绍了详细的级联过程。依据提取的OBS考虑优先级下突发包碎片特征,通过级联分类器对突发包碎片进行检测。实验结果表明,所提方法检测精度高。
OBS environment is more complicated. In order to reduce the packet loss rate and reduce the probability of network congestion and burst collision, an OBS considered packet burst detection method with priority is proposed. The OBS basic principle and OBS considered priority are introduced In the fragmentation transmission process, when the burst encounters a collision in the core node, the fragmentation is selected according to the priority of the burst and the destination core node transmits the feedback information to the edge aggregation node. The self-similar model was constructed and the statistical properties of the fractal Brownian motion were introduced. Combined with the statistical properties of the self-similar model and the fractal Brownian motion, the burst fragment features in OBS were extracted. Select Gentle Adaboost (GAB) classification algorithm, given a weak classifier, introduced a detailed cascade process. According to the extracted OBS, the priority of the burst packet fragments is considered, and the burst fragments are detected by the cascade classifier. The experimental results show that the proposed method has high detection precision.