基于平衡理論的分布式P2P信任網(wǎng)絡發(fā)現(xiàn)算法研究
發(fā)布時間：2018-10-19 10:09
【摘要】：在近幾年來P2P網(wǎng)絡技術有著飛速的發(fā)展,日常網(wǎng)絡中的多種應用以離不開P2P技術的支持。P2P網(wǎng)絡與傳統(tǒng)的C/S模式相比具有對等性、開放性等特點,給各個領域應用發(fā)展帶來了極大的便利,也正是由于這些特點,使得P2P網(wǎng)絡面臨著一些前所未有的安全問題。如惡意節(jié)點攻擊,共謀團體欺騙,惡意病毒的傳播以及知識產權保護等問題。因此如何建立P2P網(wǎng)絡的信任模型、檢測網(wǎng)絡中的惡意節(jié)點以及進行一系列的信任推測對P2P技術的未來發(fā)展具有重要的理論意義和實際價值。一些學者提出了許多根據(jù)節(jié)點歷史通訊信息或全局節(jié)點信譽值等方式構建可信的P2P網(wǎng)絡模型的方案,但許多方案在抵抗惡意節(jié)點攻擊、檢測共謀團體惡意行為以及網(wǎng)絡通信量等方面還存在著較多的問題。本文在閱讀大量相關文獻的基礎上,針對目前多種信任模型進行了具體的分析。并對目前建模算法普遍面臨的惡意行為攻擊等問題,提出了一種基于平衡理論的分布式P2P網(wǎng)絡的信任模型發(fā)現(xiàn)算法以及基于信任結構的信任發(fā)現(xiàn)算法。經過研究發(fā)現(xiàn),平衡理論中所定義的實體之間的平衡狀態(tài)可以對應到P2P網(wǎng)絡中的節(jié)點信任關系。基于平衡理論的信任建模對原始P2P網(wǎng)絡進行分析,尋找其中滿足平衡條件的三角形結構,進而將網(wǎng)絡中的所有處于平衡狀態(tài)的三角形結構提取出來,構建初步的信任網(wǎng)絡。為了解決抵制單個惡意節(jié)點的攻擊以及共謀團體的惡意行為的問題,本文在深入研究大量文獻的基礎上,提出了針對惡意節(jié)點以及共謀團體的檢測方案。在共謀團體惡意行為的檢測中,引入了節(jié)點的相關性概念,通過計算節(jié)點間的相關性進而檢測惡意共謀團體。在信任推測算法中,引入了基于平衡理論的平衡因子,分別對節(jié)點、鏈接以及整個網(wǎng)絡的平衡性進行了計算。通過假設對目標節(jié)點的信任值,評估該假設對節(jié)點的平衡性的影響,進而根據(jù)影響情況對假設的信任值進行評價。本文首次將平衡理論應用到了P2P網(wǎng)絡信任網(wǎng)絡的發(fā)現(xiàn)中,并使用Epinions數(shù)據(jù)集對模型進行了實驗分析。實驗表明基于平衡理論的信任發(fā)現(xiàn)算法以及基于平衡結構的信任發(fā)現(xiàn)算法有較好表現(xiàn),對構建P2P信任網(wǎng)絡具有重要作用。
[Abstract]:In recent years, P2P network technology has been developing rapidly. Many applications in daily network are supported by P2P technology. Compared with the traditional C / S model, P2P network has the characteristics of reciprocity and openness. It brings great convenience to the development of applications in various fields, and it is precisely because of these characteristics that P2P networks are faced with some unprecedented security problems. Such issues as malicious node attacks, collusive group spoofing, spread of malicious viruses, and intellectual property protection and so on. Therefore, how to establish a trust model of P2P network, detect malicious nodes in the network and carry out a series of trust speculation have important theoretical and practical value for the future development of P2P technology. Some scholars have proposed many schemes to construct a trusted P2P network model based on the historical communication information of the node or the reputation value of the global node, but many of the schemes are resisting malicious node attacks. There are still many problems in detecting malicious behavior of collusion groups and network traffic. On the basis of reading a lot of relevant literature, this paper makes a detailed analysis of various trust models. A trust model discovery algorithm based on balance theory and a trust discovery algorithm based on trust structure are proposed to solve the problems of malicious behavior attacks which are commonly faced by the current modeling algorithms. It is found that the equilibrium state between entities defined in equilibrium theory can correspond to the trust relationship between nodes in P2P networks. Trust modeling based on equilibrium theory analyzes the original P2P network, finds out the triangle structure which satisfies the equilibrium condition, and then extracts all the triangle structures in the equilibrium state to construct the primary trust network. In order to solve the problem of resisting the attack of a single malicious node and the malicious behavior of the collusion group, this paper proposes a detection scheme for the malicious node and the collusion group on the basis of in-depth study of a large number of literatures. In the detection of malicious acts of collusion groups, the concept of node correlation is introduced, and then the malicious collusion group is detected by calculating the correlation between nodes. In the trust conjecture algorithm, the equilibrium factor based on equilibrium theory is introduced, and the equilibrium of nodes, links and the whole network is calculated respectively. Based on the assumption of the trust value of the target node, the influence of the assumption on the balance of the node is evaluated, and then the trust value of the hypothesis is evaluated according to the influence. In this paper, the equilibrium theory is applied to the discovery of P2P trust network for the first time, and the Epinions data set is used to analyze the model. Experiments show that trust discovery algorithm based on equilibrium theory and trust discovery algorithm based on balanced structure have good performance and play an important role in building P2P trust network.
【學位授予單位】：東北大學
【學位級別】：碩士
【學位授予年份】：2013
【分類號】：TP393.08
，
本文編號：2280815
[Abstract]:In recent years, P2P network technology has been developing rapidly. Many applications in daily network are supported by P2P technology. Compared with the traditional C / S model, P2P network has the characteristics of reciprocity and openness. It brings great convenience to the development of applications in various fields, and it is precisely because of these characteristics that P2P networks are faced with some unprecedented security problems. Such issues as malicious node attacks, collusive group spoofing, spread of malicious viruses, and intellectual property protection and so on. Therefore, how to establish a trust model of P2P network, detect malicious nodes in the network and carry out a series of trust speculation have important theoretical and practical value for the future development of P2P technology. Some scholars have proposed many schemes to construct a trusted P2P network model based on the historical communication information of the node or the reputation value of the global node, but many of the schemes are resisting malicious node attacks. There are still many problems in detecting malicious behavior of collusion groups and network traffic. On the basis of reading a lot of relevant literature, this paper makes a detailed analysis of various trust models. A trust model discovery algorithm based on balance theory and a trust discovery algorithm based on trust structure are proposed to solve the problems of malicious behavior attacks which are commonly faced by the current modeling algorithms. It is found that the equilibrium state between entities defined in equilibrium theory can correspond to the trust relationship between nodes in P2P networks. Trust modeling based on equilibrium theory analyzes the original P2P network, finds out the triangle structure which satisfies the equilibrium condition, and then extracts all the triangle structures in the equilibrium state to construct the primary trust network. In order to solve the problem of resisting the attack of a single malicious node and the malicious behavior of the collusion group, this paper proposes a detection scheme for the malicious node and the collusion group on the basis of in-depth study of a large number of literatures. In the detection of malicious acts of collusion groups, the concept of node correlation is introduced, and then the malicious collusion group is detected by calculating the correlation between nodes. In the trust conjecture algorithm, the equilibrium factor based on equilibrium theory is introduced, and the equilibrium of nodes, links and the whole network is calculated respectively. Based on the assumption of the trust value of the target node, the influence of the assumption on the balance of the node is evaluated, and then the trust value of the hypothesis is evaluated according to the influence. In this paper, the equilibrium theory is applied to the discovery of P2P trust network for the first time, and the Epinions data set is used to analyze the model. Experiments show that trust discovery algorithm based on equilibrium theory and trust discovery algorithm based on balanced structure have good performance and play an important role in building P2P trust network.
【學位授予單位】：東北大學
【學位級別】：碩士
【學位授予年份】：2013
【分類號】：TP393.08
，
本文編號：2280815
本文鏈接：http://sikaile.net/falvlunwen/zhishichanquanfa/2280815.html
