公开(公告)号：US06785240B1

公开(公告)日：2004-08-31

申请号：US09585738

申请日：2000-06-02

申请人： Jin Cao ,  R. Drew Davis ,  Scott Alan Vander Wiel ,  Bin Yu

发明人： Jin Cao ,  R. Drew Davis ,  Scott Alan Vander Wiel ,  Bin Yu

IPC分类号： H04J116

CPC分类号： H04L43/067 ,  H04L41/0213 ,  H04L41/142 ,  H04L41/16 ,  H04L43/045 ,  H04L43/0888

摘要： In many packetized communication networks, it is not feasible to obtain exact counts of traffic (OD counts) between specific origin-destination node pairs, because the link counts that are readily obtainable at router interfaces are aggregated indiscriminately over OD pairs. The best that can be done is to make a probabilistic inference concerning the OD counts from the observed link counts. Such an inference relies upon a known linear relationship between observed link counts and unknown OD counts, and a statistical model describing how the values of the OD and link counts are probabilistically distributed. Disclosed is an improved method for making such inferences. The disclosed method takes explicit account of past data when forming a current estimate of the OD counts. As a consequence, behavior that evolves in time is described with improved accuracy and smoothness.

摘要翻译： 在许多打包通信网络中，在特定的起始 - 目的地节点对之间获得精确的业务计数（OD计数）是不可行的，因为在路由器接口上容易获得的链路计数在OD对上不加区分地聚合。 可以做的最好的事情是从观察到的链接数量中对OD计数进行概率推断。 这样的推论依赖于观察到的链接计数和未知OD计数之间已知的线性关系，以及描述OD和链接计数值如何概率分布的统计模型。 公开了一种用于进行这种推断的改进方法。 当形成OD计数的当前估计时，所公开的方法明确地考虑过去数据。 因此，在时间上演变的行为以提高的准确性和平滑性来描述。