论文部分内容阅读
现有提供确定性性能保证的核心无状态分组调度策略中,可扩展性的获得通常以中间节点的速率损失为代价,从而导致数据传输突发性的增强,降低了网络资源的利用率。为了减少速率损失,提出一个核心无状态调度框架DVRS(de lay vector reference system),通过运用虚拟延迟矢量技术,在保证端到端延迟上限与V irtua lC lock算法相同的同时,将核心无状态算法的基本速率保证能力从现在的首尾点精确保证提高到逐点精确保证;并利用局部平均法满足用户对处理开销与带宽保证精度的不同需求;另外,由于核心节点操作的相似性,DVRS类算法具有较好的配置渐进性。
In the existing state-of-the-art packet scheduling strategy that provides deterministic performance guarantees, scalability is usually obtained at the expense of the rate loss at the intermediate nodes, resulting in a sudden increase in data transmission and a reduction in the utilization of network resources. In order to reduce the rate loss, a core stateless scheduling framework DVRS (de lay vector reference system) is proposed. By using the virtual delay vector technique, while guaranteeing the same end-to-end delay limit as the Viteru lC lock algorithm, the core stateless algorithm The basic rate guarantee ability can be improved from accurate assurance at the beginning and end of the current period to precise assurance at the point by point. The local average method can meet the users’ different requirements on processing overhead and bandwidth assurance accuracy. In addition, due to the similarities of core node operations, Has a good configuration of progressive.