唐明董博士近期在SCIENCE CHINA Information Sciences发表题为“Integrating local and partial network view for routing on scale-free network”学术论文一篇。唐明董博士现为湖南科技大学计算机科学与工程学院副教授、硕士生导师,计算机科学与技术系系主任。2010年1月博士毕业于中国科学院计算技术研究所。2012年1月至今,以博士后身份公派至美国密苏里科技大学深造。
===============================================================================================
标题: Integrating local and partial network view for routing on scale-free network
作者: MingDong Tang; GuoQiang Zhang; Yi Sun; JianXun Liu; Jing Yang; Tao Lin
来源出版物:Science China Information Sciences
DOI:10.1007/s11432-012-4655-y 出版年: Aug.2012
Print ISSN:1674-733X Online ISSN: 1869-1919
摘要:Traditional routing schemes, such as OSPF, optimize data plane routing efficiency by maintaining full view of the network at the control plane. However, maintaining full network view and handling frequent routing information updates are costly in large-scale complex networks, which are considered to be the root causes for the routing scalability issue. Recently, it is suggested that routing on local or partial information is plausible if slight performance degradation is acceptable. This paper proposes a routing scheme, operating on an integrated network view at each node that consists of its local neighborhood and a globally unique skeleton tree. This scheme significantly reduces storage, communication and processing costs. On scale-free networks, this benefit only comes at the cost of marginal performance degradation, which implies that it is not worthwhile to do shortest path routing based on full view of the network on scale-free networks. In contrast, the routing efficiency is severely aggravated on purely random networks, indicating the inappropriateness of this scheme and the rationality of maintaining full network view on random networks.
(彭桃供稿)