机会网络中自私行为影响的分析与调节

以往针对机会网络中节点自私性影响评估的研究,建立了多种分析模型,然而并未考虑如何应用这些理论模型.本文分别在网络无TTL限制和有TTL限制两种情况下,利用评估模型对自私节点的影响进行分析,并提出一种基于消息分类的自私模型,评估了自私节点对延迟和消息投递率的影响.分析表明,机会网络受自私节点的影响不仅与自私节点的数目有关,而且与节点的移动速度,消息传输范围有关.因此,可以通过调节这些参数提高网络的性能,减少自私节点的影响.首先,通过理论推导分别得出了2-hop路由和传染病路由在延迟保证下的合作节点数目和消息传输范围以及传染病路由在消息投递率保证下消息的TTL.随后,通过ONE仿真平台对自私节点影响的评估,证明自私行为影响下网络性能的理论值与实验值近似,并且证明了通过理论分析获得的参数值也与实验值近似.     

机会网络中能量有效的副本分布状态感知路由机制

针对机会网络中的节点能耗与消息扩散程度之间存在矛盾,基于节点能耗与副本状态感知的路由机制可以显著地改善资源受限场景的网络性能,综合考虑消息扩散程度与节点剩余能量,并结合节点相遇概率预测方法,提出能量有效的副本分布状态感知路由机制,为消息分布式地选择合理的中继节点。仿真结果表明,所提出的路由机制能够有效降低网络负载,并大幅改善消息投递率和平均时延性能。     

感知无线电自私网络低开销诚实能效路由

在感知无线电自私ad hoc网络环境下提出一种低代销的诚实能效路由(Lowest Cost Path,LCP),将路由的开销结合对主用户的干扰和节点剩余能量,利用格罗夫斯-克拉克-威科瑞(Vickrey-Clark-Groves,VCG)定价机制给LCP上每个中继节点以一定报酬促使节点之间的合作,并证明VCG机制下节点真实的反映自身的类型是一个占优策略。仿真结果表明,该路由算法有效实现了感知无线电自私ad hoc网络的诚实能效路由,与现有最短路径路由协议比较,该协议具有更好的性能。     

基于连接质量和节点相关度的机会网络路由

提出利用连接质量估计节点间消息成功转发的概率,并证明了节点间相关度越高,消息转发率越高。在此基础上,提出基于连接质量和节点相关度的机会网络路由机制。路由计算节点转发效用值时,综合了连接质量和节点相关度;消息转发时,不断向相对于目的节点转发效用值高的节点转发消息。仿真结果表明,与传统路由相比,在消息送达率相同的前提下,该路由消息延迟更小、消耗的总能量更少。     

一种不可信环境下的匿名位置辅助路由激励机制

不可信环境下的路由性能问题,是移动自组织网(mobile ad-hoc networks,MANET)匿名路由协议面临的主要问题之一.本文为在不可信环境下通过对自私节点进行高效的协作激励,提高匿名路由协议的性能,提出了一种基于哈希链的匿名位置辅助路由激励机制.该机制利用哈希链在计算上的高效性和安全上的不可逆性,达成了对匿名数据转发节点的即时激励;通过基于支付代价的路由选择机制,优化了现有位置辅助路由机制的路由发现过程.匿名性分析证明,该机制能保证参与路由节点的匿名性.效率评价表明,在数据传输总量较大时,该机制对路由性能的影响很小,且该机制在较小规模的网络中有更好的性能.     

社区结构感知的间断连接无线网络路由机制

针对间断连接无线网络中消息投递成功率低的问题1依据网络中节点社会关系的差异性,提出了一种社区结构感知的路由机制.该机制根据节点在各个运动周期内的状态信息建立马尔可夫模型,以描述节点运动状态转换过程,进而以分布式的方式感知节点中心度,并以社区中心度为参数,采用社区标签交换方法对网络结构进行动态检测,最终利用社区内中心节点为中继辅助完成消息的转发.仿真结果表明,所提出的路由机制在投递率方面的性能改善程度接近90％,极大地优化了网络性能.     

一种基于邻居合作监测的移动Ad hoc网络信任模型

针对自私节点的恶意丢包行为,将节点和其邻居节点所监测的结果结合起来,计算出节点间的信任度,并以此作为路由选择的依据来促进节点间的相互合作,提出一种基于邻居合作监测的移动Adhoc网络信任模型.将信任模型应用于DSR路由协议并在NS2中进行仿真实验,结果表明该信任模型可以有效地缓解自私节点造成的影响,提高了网络的分组投递率.     

一种解决ad hoc网络中自私节点安全问题的方法

移动ad hoc网络（MANET）路由和分组转发功能需要节点间的相互合作，自私节点为节约能量而不参与合作，将会影响网络的正常工作．因此本文将针对自私节点危害和解决方法展开分析，并提出一种改进解决方法——基于可信任中心服务器和邻居监测的合作增强机制，该方法不但能够激励节点间的相互合作，而且有效地解决了自私节点改变用户身份的欺骗攻击．最后，仿真结果表明这种合作增强机制对自私节点安全问题是有效的，并且易于实用化．     

DTN网络中基于生灭模型的节点运动模式检测

提出了一种DTN网络中基于生灭模型的节点运动模式检测方法(MMD-BDM, mobility model detection method based on birth and death model)。在节点数据传输过程中,根据数据传输量计算得到各个连通链路采样时刻的信道容量,在本节点构建以信道容量构成的三维时变连通矩阵,并在此基础上推导出基于边连通的节点连通性生灭模型,然后分析节点连通边的生灭特征并以此判断节点的运动模式,以优化DTN网络消息投递过程中的路由转发策略。最后,将该运动模式检测方法应用在喷射路由(SWR, spray and wait routing)和随机网络编码路由(RNCR, random network coding routing)算法中进行了仿真实验,实验表明,该方法能提高DTN网络机会路由的投递率和投递延迟等数据投递性能。     

基于社会特征周期演化的机会移动网络路由转发策略

针对分布式k团社区检测引起的超大社区问题,提出了具有节点退出机制的?-window社区检测方法,相应提出了?-window中心性估计。通过实验发现?-window社区和?-window中心性具有周期演化特性,利用该特性,提出TTL(time to live)社区检测和TTL中心性估计,以更准确预测消息生存期上节点的相遇。随后,利用TTL社区和TTL中心性作为转发测度,设计了新的机会移动网络路由算法PerEvo。实验结果表明,与现有的基于社会特征的路由算法比较,PerEvo在保持基本不变的传输开销的同时,有效提高了机会移动网络消息投递的成功率。 关键词：机会移动网络;社区;中心性;周期演化;消息转发     

Exploiting Selfishness,Altruism and Common Welfare to Enhance Performance of Routing Protocols for Mobile Ad hoc Networks

Mobile Ad hoc Networks are sensitive to selfish behavior that may occur due to restricted power or other resources. Several approaches have been investigated so far to address this problem. In many of them, upon detection, a selfish node is punished with isolation from network services access and in most cases with no possibility for redemption. In this paper, we show that selfish behavior can be exploited to improve network performance. We modify an existing routing protocol by introducing an altruism coefficient to model the overall satisfaction of every node from the network services. When the altruism coefficient is increased, the selfish behavior is decreased. We extend our approach by introducing a common welfare factor, which actually enforces the nodes to cooperate when the network connectivity is critical. A network simulator is utilized to show the impact of our schemes on the performance of the routing protocol when selfish nodes are present in the network.

     

ICARUS: hybrId inCentive mechAnism for coopeRation stimUlation in ad hoc networkS

Mobile ad hoc networks rely on the cooperation of nodes for routing and forwarding. However, individual nodes may not always be willing to cooperate. In order thus to stimulate cooperation in ad hoc networks, several incentive mechanisms have been developed. In this paper we propose a new hybrid incentive mechanism, called ICARUS, which is an extension of DARWIN, a well-known reputation-based mechanism, combining advantages of both reputation-based and credit-based mechanisms. The objective of ICARUS is to detect and punish selfish nodes efficiently and at the same time motivate nodes to cooperate by rewarding the packet forwarding. Furthermore, ICARUS ensures fairness for distant nodes and prevents selfish nodes from corrupting the system using false information. The proposed scheme’s performance is tested through extended series of simulations and is compared with DARWIN. We show that ICARUS detects and isolates selfish nodes much faster, while at the same time improves the Quality of Service (QoS) received by non-selfish nodes, including distant ones.     

OMH—Suppressing Selfish Behavior in Ad hoc Networks with One More Hop

In ad hoc networks, wireless nodes rely on each other to transmit data over multi-hops by forwarding packets. A selfish node may decide not to forward packets for other nodes to save its own resource but still use the network to send and receive data. Such a selfish behavior can degrade network performance significantly. Most existing work took observation, reputation and token based mechanisms. However observation based mechanism suffers from mobility and collusion; reputation and token based mechanisms suffer from system complexity and efficiency. In this paper, we propose One More Hop (OMH) protocol which suppresses selfish behavior from a totally new angle. Basing on the fact that the selfish but rational nodes still want to receive and send packets, if a node can not determine whether a packet is destined for it or not, it can not drop the packet. With modified routing protocol and cryptographic techniques, OMH achieves this design target. It is robust and efficient. The simulation shows that OMH works well under different network situations.     

基于博弈的协作路由算法

协作虚拟多输入多输出（VMIMO）传输是一种有效的无线传输性能优化技术。将物理层协作VMIMO技术和网络层路由选择技术相结合,设计跨层VMIMO路由选择方案可以利用VMIMO的分集增益,显著地降低网络传输能耗。如何设计VMIMO协作路由协议抵抗无线网络的自私节点和欺骗行为,保证高数据转发率和低传输能耗成为路由设计中的重大挑战。为了提高自私网络的VMIMO路由性能,提出了一种基于重复路由博弈的VMIMO协作路由算法。该算法将网络划分成多个Group、Group间使用VMIMO传输数据。将Group间路由选择过程建模为重复路由博弈过程。为了提高数据转发的成功率,提出适用度函数评估节点参与数据分组转发的信誉。以此为基础,提出基于适用度的路由选择子算法和路由转发子算法。理论证明所提重复路由博弈可达到帕累托最优。仿真实验结果表明本算法可以促进自私节点相互合作,可获得较高的数据转发率,较好地减少数据传输时延以及能量消耗。     

Incentive based scheme for improving data availability in vehicular ad-hoc networks

Vehicular networks are popular in recent years to provide low cost communication medium during mobility. Vehicular Delay Tolerant Networks (DTNs) are one of the major categories of emerging technology. DTNs work on carry and forward mechanism to deliver data to the destination. The network performance gets severely affected due to reluctance shown by selfish nodes where few nodes show no interest in forwarding others data due to lack of any personal profit. The proposed mechanism is based on coalition game theory and discusses about incentive based mechanism which provides incentive to nodes which are forwarding data to forward to destination and motivates other vehicles in the network to participate in coalition to forward data. This scheme not only encourages other selfish nodes to forward their private data and other nodes’ public data as early as possible to destination but also increases reliability in the network as more nodes show their interest in selected routing protocol. The proposed scheme outperforms in overall benefit earned by individual node and whole coalition, and increases mutual cooperation which improves availability of data in the network.     

Performance modeling of DTN routing with heterogeneous and selfish nodes

The performance modeling study of Delay-Tolerant Network routing, in general, assumes the nodes to be homogeneous (in terms of features such as the coverage range) and uncompromised (in terms of forwarding messages). However, in realistic settings this may not be the case. The routing performance modeling of such realistic scenarios that involve multifariously-featured and egotistic nodes would be interesting and insightful. To this end, in this paper, we analytically model the routing behavior of such nodes using Ordinary Differential Equations for two different routing protocols namely, Epidemic Routing and Two-Hop Routing. Furthermore, we also study the degradation in the routing performance caused by an increase in the fraction of selfish nodes present in the heterogeneous node population. The proposed analytical model is validated via extensive simulations.     

基于信誉的延迟容忍网络激励方案

由于长延迟、频繁中断等特性,延迟容忍网络采用“存储-携带-转发”的路由方式实现报文的传递.这种传输方式建立在网络节点互相信任,诚实合作的假设之上.然而,由于受到自身资源的限制,网络中存在自私节点拒绝参与网络合作的行为,导致网络传输性能下降,甚至无法正常运行.针对此类问题,本文提出一种基于信誉的延迟容忍网络激励方案,鼓励节点共享资源,并约束自私行为.设计了一种观测协议,以解决延迟容忍网络环境下对网络节点行为的观测问题;提出了一种信誉模型,通过信任值评价节点的可信程度,从而优化路由决策,惩罚自私节点.仿真结果表明了本文提出的激励方案的有效性和可行性.