手机安全支付异常流量监测技术的研究

发布时间：2018-04-25 14:15

本文选题：手机支付 + 安全监测　；参考：《北京交通大学》2014年硕士论文

【摘要】：摘要：近年来,手机移动支付因其便利性和快捷性具有了广大的市场,但因其应用层的安全机制和技术还不够完善,外加手机病毒的出现、移动终端与银行接口以及系统的开放性等方面存在的漏洞,使得手机支付的安全性成为当前的研究热点之一。本论文主要以研究手机支付中恶意行为带来的网络流量异常变化的安全威胁问题为切入点,以网络流量的分析提取为基础,以信息熵为基本度量方法,建立了基于多层前馈神经网络(Back Propagation Neural Network,简称BP神经网络)的手机安全支付异常流量监测模型。系统通过对网络流量变化过程中异常行为的提取和对比,给出了一个完整的从异常发现到异常分析及判断异常指标动态调整的工作流程。本文所作的工作主要体现在以下几个方面： 1.通过对手机移动支付过程的分析,讨论了其安全威胁的主要来源,并选取了特征参数对其流量进行监测从而达到保证其安全性的目的。其中,通过分析手机支付中能够充分体现其流量特征的参数,提取出四个特征参数并计算熵值。 2.详细设计了基于信息熵和BP神经网络的异常流量监测系统的结构。利用BP神经网络最突出的再学习机制,将初步监测结果再次送入神经网络内部,达到动态分析和调整的目的。 3.提出了动态调整网络流量阈值区间的算法,通过该算法可以实时更新流量监测指标,做到适应当前网络状况的要求。此外,考虑到监测任务中监测结果的效率问题,设计该算法时只选取变化显著的特征参数参与调整,从而减少了网内数据的存储量和计算量。本论文最后利用matlab实现了本文提出的异常流量监测模型,并对系统进行了测试和分析。结果表明：该模型较为成熟有效,能够起到一定的抵制网络恶意攻击的作用,适用于手机安全支付系统异常流量监测。
[Abstract]:Absrtact: in recent years, mobile payment of mobile phone has a broad market because of its convenience and quickness, but the security mechanism and technology of its application layer are not perfect enough, and the emergence of mobile phone virus. The security of mobile payment has become one of the current research hotspots due to the vulnerabilities in the interface between mobile terminal and bank and the openness of the system. In this paper, we focus on the research of the security threat caused by malicious behavior in mobile phone payment, based on the analysis and extraction of network traffic, and based on the information entropy as the basic measurement method. A mobile phone security payment abnormal traffic monitoring model based on the multilayer feedforward neural network back Propagation Neural network (BP neural network) is established. Through the extraction and comparison of abnormal behavior in the process of network traffic change, a complete workflow from anomaly detection to anomaly analysis and dynamic adjustment of abnormal index is presented. The work done in this paper is mainly reflected in the following aspects: 1. By analyzing the mobile payment process of mobile phone, the main sources of security threat are discussed, and the characteristic parameters are selected to monitor the traffic to ensure the security of mobile phone. Among them, by analyzing the parameters of mobile phone payment which can fully reflect the characteristics of its flow, four feature parameters are extracted and entropy is calculated. 2. The structure of abnormal flow monitoring system based on information entropy and BP neural network is designed in detail. Using the most outstanding relearning mechanism of BP neural network, the preliminary monitoring results are sent into the neural network again to achieve the purpose of dynamic analysis and adjustment. 3. An algorithm for dynamically adjusting the threshold interval of network traffic is proposed, through which the traffic monitoring index can be updated in real time to meet the requirements of current network conditions. In addition, considering the efficiency of the monitoring results in the monitoring task, only the characteristic parameters with significant changes are selected in the design of the algorithm, which reduces the storage and computation of the data in the network. In the end of this paper, the matlab is used to realize the model of abnormal flow monitoring, and the system is tested and analyzed. The results show that the model is more mature and effective and can resist network malicious attacks to some extent. It is suitable for mobile phone security payment system to monitor abnormal traffic.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.53

【参考文献】