基于时延—带宽约束的新型层次拓扑聚集算法

Delay-bandwidth constrained topology aggregation algorithm

针对时延—带宽约束的层次路由问题,提出了楼梯法这一全新的拓扑聚集算法。使用一个规则的楼梯来模拟2个边界节点之间的所有通路支持的QoS属性,将原始拓扑压缩为一个六元组标识的全相连图。然后,根据六元组各参数特性对全相连图分别采用了最大生成树和改进的星型压缩算法进行线性压缩。仿真的结果表明,在聚集后信息空间复杂度同为O(|B|)的情况下,楼梯法较其他拓扑聚集算法具有更小的信息失真率。

In order to deal with hierarchical routing with delay and bandwidth constraints, a new topology aggregation algorithm called stair method was proposed. A regular virtual stair was used to represent the QoS property of all paths between any two border nodes. Then a full mesh with each link denoted by six tuples was constructed. According to the different characters of the parameters of six-meshes, the maximum weight spanning tree and star structures were adopted to compress them respectively. Simulation results show stair algorithm achieves not only better performance in terms of the loss ratio of state information of the network, but also the same space complexity of O（｜B｜） with the existing algorithms.