Details of Research Outputs

提出一种可扩展、高效的树结构应用层组播协议——FTALM,通过将组播组成员构成的列表进行递归分割,构造一棵可高效传送组播数据的层次化斐波那契树.引入新的概念"本地区域",通过考虑底层的拓扑结构特点来减少在高代价链路上的数据传输.在构树过程中,考虑本地区域的成员数目以获得一棵平衡的树结构.底层拓扑结构的考虑以及斐波那契树的构造使FTALM的组播延迟性能得到提高.模拟实验结果证实了FTALM在组播延迟性能上的优势,在大型组播组的情况下优势更强. 

A scalable and efficient tree structured application-layer multicast protocol-FTALM (Fibonacci tree application-layer multicast) was proposed. By recursively partitioning the member sequences into two halves with different lengths, a hierarchical Fibonacci tree was built to multicast packets efficiently. In FTALM, the underlying topology characteristic was considered by introducing an idea of local area to reduce packet delivering on costly links. Moreover, the size of each local area was taken into account when constructing the hierarchical architecture in order to obtain a balanced architecture. The consideration of underlying topology and the construction of Fibonacci tree improved the delay performance of FTALM. The simulation results show the benefits of FTALM and demonstrate that it works well especially for large multicast groups.