首页> 外文会议>Global Telecommunications Conference >A quantitative study of differentiated services for the Internet
【24h】

A quantitative study of differentiated services for the Internet

机译:对互联网差异化服务的定量研究

获取原文

摘要

The differentiated services architecture provides router mechanisms for aggregate traffic, and edge mechanisms for individual flows, that together can be used to build services with varying delay and loss behavior. We compare the loss and delay behavior that can be provided using the services based on combinations of two router mechanisms, threshold dropping and priority scheduling and two packet marking mechanisms, edge-discarding and edge-marking. We compare the delay and loss behavior of the two router mechanisms coupled with edge-discarding for a wide range of traffic arrivals. We observe that priority scheduling provides lower expected delays to preferred traffic than threshold dropping. In addition, we find that a considerable additional link bandwidth is needed with threshold dropping to provide same delay behavior as priority scheduling. We further observe little difference in the loss incurred by preferred traffic under both router mechanisms, except when sources are extremely bursty, in which case threshold dropping performs better. We examine the throughput of a TCP connection that uses a service built upon threshold dropping and edge-marking. Our analysis shows that a significant improvement in throughput can be achieved. However, we find that in order to fully achieve the benefit of such a packet marking, the TCP window must take the edge-marking mechanism into consideration.
机译:差异化服务体系结构为聚合流量提供路由器机制,以及各个流的边缘机制,可以用于构建具有不同延迟和损耗行为的服务。我们比较丢失和延迟行为,这些丢失和延迟行为可以根据两个路由器机制,阈值下降和优先级调度以及两个分组标记机制,边缘丢弃和边缘标记的组合提供。我们比较两个路由器机制的延迟和损失行为,耦合到边缘丢弃的广泛交通到达。我们观察到优先调度为优先的流量提供更低的预期延迟而不是阈值下降。此外,我们发现,需要相当大的额外的链路带宽,阈值下降以提供与优先级调度相同的延迟行为。除了源极度突发时,我们进一步观察到在两个路由器机制下所产生的优选流量所产生的损失差异很小,在这种情况下,阈值下降表现更好。我们检查使用基于阈值下降和边缘标记的服务的TCP连接的吞吐量。我们的分析表明,可以实现吞吐量的显着改善。但是,我们发现,为了充分实现这种数据包标记的好处,TCP窗口必须考虑边缘标记机制。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号