一种软件定义网络中基于博弈的混合路由算法*

现存软件定义网络中的流量工程算法考虑单一,难以实现流量均衡,且缺乏对端到端的时延优化。对此提出一种博弈混合路由算法,以少量的显式路由与基于目的的聚合流量配置,在保证流量均衡的同时,优化网络的平均端对端时延。聚合流量配置采用纳什议价博弈理论对流量均衡与端对端时延优化场景进行联合优化建模,并设计了相应的启发式算法进行求解。实验结果表明,与现有算法相比,流量均衡程度和平均端对端时延均得到一定程度的改善,在Splintlink拓扑中,最大链路利用率平均减少6%,平均端对端时延减少12%。     

A secure routing scheme based on social network analysis in wireless mesh networks无线网状网中一种基于社会网络分析的安全路由机制

As an extension of wireless ad hoc and sensor networks, wireless mesh networks (WMNs) are employed as an emerging key solution for wireless broadband connectivity improvement. Due to the lack of physical security guarantees, WMNs are susceptible to various kinds of attack. In this paper, we focus on node social selfish attack, which decreases network performance significantly. Since this type of attack is not obvious to detect, we propose a security routing scheme based on social network and reputation evaluation to solve this attack issue. First, we present a dynamic reputation model to evaluate a node’s routing behavior, from which we can identify selfish attacks and selfish nodes. Furthermore, a social characteristic evaluation model is studied to evaluate the social relationship among nodes. Groups are built based on the similarity of node social status and we can get a secure routing based on these social groups of nodes. In addition, in our scheme, nodes are encouraged to enter into multiple groups and friend nodes are recommended to join into groups to reduce the possibility of isolated nodes. Simulation results demonstrate that our scheme is able to reflect node security status, and routings are chosen and adjusted according to security status timely and accurately so that the safety and reliability of routing are improved.