网络业务流行为的建模和分析

发布时间：2018-03-03 08:07

本文选题：业务流行为　切入点：行为建模　出处：《北京邮电大学》2014年硕士论文　论文类型：学位论文

【摘要】：近年来,随着用户数目的增加,网络规模成指数增长；与此同时,网络所承载的业务类型也呈现出多样化,综合化的趋势。这种大规模结构不仅造成网络测量的困难,而且由于大量不同业务的相互影响导致网络状态不断发生变化,要想精确把握网络的整体性能,了解网络资源的“去向”,至今缺乏行之有效的解决途径。在复杂的网络环境下,如何满足用户的需求和提高网络性能,前人进行了极具价值的研究和探索,主要基于以下几个方面：一是从业务流分布的宏观特征出发,寻找一种基于一个或者多个目标的网络资源分配最优方式；二是对整体网络拓扑的研究,探索网络拓扑对网络性能的影响；三是依据网络外部的特征参数,提出网络性能评价的模型。但是,仅仅从宏观的角度来理解整个网络的运作,我们认为这是不充分的,从微观来看,网络的复杂性在一定程度上也来源于业务的相关性。以此为出发点,综合考虑网络拓扑对数据流行为的影响以及数据流自身需求的需要,我们提出建立不同需求和属性的业务网络业务流的基本模型,而整个网络的行为正是由这些基本模型相互作用和影响的结果。这种相关性我们称之为耦合行为。研究表明,在整个网络的范围内,耦合行为导致了业务流行具有扩散、波动和叠加的性质,这使得整个网络性能在已经拥塞的情况下(包括具有反馈机制的策略和无反馈机制的策略)进一步恶化,即“马太效应”。此外,我们对这种行为在两种常见的网络传输策略下进行了分析,得到了最接近实际网络中的情形。进一步,我们将模型扩展到网络能量,提出了网络势能的概念来分析整个网络性能。
[Abstract]:In recent years, with the increase in the number of users, the scale of the network has grown exponentially. At the same time, the types of services carried by the network have shown a trend of diversification and integration. Moreover, due to the mutual influence of a large number of different services, the status of the network is constantly changing. To accurately grasp the overall performance of the network and to understand the "destination" of network resources, there is a lack of effective solutions. In the complex network environment, how to meet the needs of users and improve the performance of the network, predecessors have carried out a very valuable research and exploration, mainly based on the following aspects: first, from the macro characteristics of traffic flow distribution, To find an optimal way to allocate network resources based on one or more objectives; second, to study the overall network topology to explore the impact of network topology on network performance; third, according to the characteristics of the external parameters of the network, A model of network performance evaluation is proposed. However, we think it is not sufficient to understand the operation of the whole network from a macro perspective, and from a microscopic point of view, the complexity of the network also comes from the relevance of the business to some extent. Considering the influence of network topology on the behavior of data flow and the demand of data flow itself, we propose a basic model of business network traffic with different requirements and attributes. The behavior of the whole network is the result of the interaction and influence of these basic models. This correlation is called coupling behavior. The nature of volatility and superposition further worsens the performance of the entire network in already congested situations (including strategies with feedback mechanisms and those with no feedback mechanisms), known as the "Matthew effect." We analyze this behavior under two common network transmission strategies, and get the closest case to the real network. Further, we extend the model to network energy. The concept of network potential energy is proposed to analyze the network performance.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.06

【共引文献】