MERGING AND SPLITTING SECOND-ORDER SELF-SIMILAR PROCESSES (TRAFFICS)

Recent traffic measurements in corporate LANs, Variable Bit Rate (VBR) video sources, ISDN control channels, and other communication systems, have indicated traffic behavior of self similar nature, which has implications for design, control and analysis of high speed networks. Merging and splitting are two basic networking operations. This paper gave the necessary and sufficient conditions for that merging of second order self similar traffic streams also results in a second order self similar stream. It shows that splitting traffic streams of the second order self similar stream are still self similar streams by the independent splitting operation.

An Intrusion Alarming System Based on Self-Similarity of Network Traffic

Intrusion detection system ean make effective alarm for illegality of networkusers, which is absolutely necessarily and important to build security environment of communicationbase service According to the principle that the number of network traffic can affect the degree ofself-similar traffic, the paper investigates the variety of self-similarity resulted fromunconventional network traffic. A network traffic model based on normal behaviors of user isproposed and the Hursl parameter of this model can be calculated. By comparing the Hurst parameterof normal traffic and the self-similar parameter, we ean judge whether the network is normal or notand alarm in time.

Providing QoS guarantees for self-similar traffic flows

Provisioning network resource to meet the quality of Service (QoS) demand is a key issue for future network services. Such functions may be realized by an admission control algorithm, which determines whether or not a new traffic flow can be admitted into the network. It is widely accepted that many traffic flows have self-similar character that has detrimental influence on network performance. This characteristic has made old mathematical models invalid, and a new model must work with self-similar fractal instead. This paper applies Fractional Brownian Motion(FBM) model and integrates it into the comprehensive admission control scheme, which takes account of aggregated traffic behavior to get the statistical multiplexing performance gain. Experiment verifies that FBM model can be used to realistically describe packet traffic in modern packet networks and accurately predict their performance.

Comparative Analysis of TCP-Protocol Operation Algorithms in Self-Similar Traffic

This paper presents simulation modelling of network under the conditions of self-similar traffic and bottleneck occurrence. The comparative analysis of different TCPs (NewReno, Reno, Tahoe and etc.) has been conducted along with the testing of various algorithms of these protocols activity. The use of TCP Vegas has been proved to be the most effective.

On use of the alpha stable self-similar stochastic process to model aggregated VBR video traffic

The alpha stable self-similar stochastic process has been proved an effective model for high variable data traffic. A deep insight into some special issues and considerations on use of the process to model aggregated VBR video traffic is made. Different methods to estimate stability parameter a and self-similar parameter H are compared. Processes to generate the linear fractional stable noise （LFSN） and the alpha stable random variables are provided. Model construction and the quantitative comparisons with fractional Brown motion （FBM） and real traffic are also examined. Open problems and future directions are also given with thoughtful discussions.

基于流量自相似性的网络队列管理算法

网络流量的自相似性会导致数据突发状态持续,传统队列管理算法无法对网络流量突发状态进行预测,从而影响网络端到端时延、丢包率和吞吐性能。针对该问题,提出一种基于网络流量预测的主动队列管理算法P-ARED。基于网络流量的均值和方差给出网络流量等级的概念,讨论网络流量等级转移概率与Hurst参数之间的关系,提出基于贝叶斯估计思想的网络流量等级预测方法。在此基础上,在对自相似网络流量环境下的平均队列长度、缓存队列长度最小阈值等参数优化设置的基础上,基于Hurst参数和自相似流量等级预测结果,重新设计ARED算法中分组丢弃概率的计算方法,以提高缓存队列长度的稳定性。仿真结果表明,P-ARED算法与对比的主动队列管理算法相比,降低了网络端到端时延和丢包率,提高了端到端吞吐性能,其中平均吞吐量最高提升7.63%,平均时延最多降低17.52%。

STRUCTURAL MODELING OF WAN TRAFFIC

High resolution traffic measurements from modern communications networks provide unique opportunities for developing and validating mathematical models for aggregate traffic, or WAN (wide area network) traffic. To exploit these opportunities, this paper emphasized the need for structural model that takes into account specific physical feature of the underlying communication network structure. This approach is in sharp contrast to the traditional black box modeling methodology from this time series analysis that ignores, in general, specific physical structure. The paper demonstrated, in particular, how the proposed structural modeling approach provides a direct link between the observed self similarity characteristic of measured aggregate network traffic and the strong empirical evidence in favor of heavy tailed, infinite variance phenomena at the level of individual network connections.

基于网络流量水平等级预测的自适应随机早期检测算法

针对随机早期检测及其变体算法中平均队列长度和数据包最大丢弃概率的计算对网络流量的变化反应迟缓的问题,提出了一种基于网络流量水平等级预测的自适应随机早期检测算法。基于自相似网络流量的统计特性,建立了网络流量水平等级转移概率表,设计了自相似网络流量水平等级预测方法,该方法复杂度较低且精度较高。进一步,将预测结果应用于等间隔平均队列长度计算及数据包最大丢弃概率调整中,在固定和可变瓶颈链路容量2种情况下的仿真发现,无论自相似程度如何,所提算法在丢包率和吞吐量方面都有提升,特别在Hurst参数较大且流量负载较低时,吞吐量性能提升较大。

小时间尺度网络流量异常入侵检测算法研究

为了确保网络的安全性,提升异常入侵检测效果,提出小时间尺度网络流量异常入侵检测算法。通过改进的整体经验模态——去趋势波动分析法估计自相似指数,在大时间尺度上,利用自相似指数判断网络流量是否出现异常入侵;若存在异常入侵,则在小时间尺度上依据李氏指数和信号奇异性的相应关系,结合小波变换计算异常网络流量的李氏指数,根据李氏指数的变化情况检测网络流量异常入侵点,确定异常入侵发生时间。实验结果表明:该算法可精准判断小时间段网络流量是否存在异常入侵情况,同时可精准检测异常入侵发生时间;在不同网络流量异常入侵高度时,该算法具备较高的检测精度。

改进粒子群算法优化BP网络实现流量预测

近年来,随着网络技术的不断发展,网络传输业务也在不断增加,网络业务流量数据与日俱增。飞速增长的流量数据给网络结构和信息传输带来深刻的影响,大量的网络数据在有限的空间内进行传输,网络拥塞现象难以避免,造成网络传输时延增大、传输速率降低。因此,为了更有效地进行网络管理,改善网络传输性能,有效预测业务流量数据的变化趋势,实行具体、全面的网络管理是十分必要的。该文针对网络流量表现出的自相似性和可预测性,提出了改进PSO-BP算法,并建立其预测模型。结果显示,在网络流量预测领域内,与常规BP网络流量预测算法相比,引入改进PSO-BP算法的神经网络预测模型表现出更为优异的性能。

自相似流量的长短期记忆网络预测模型研究

随着宽带卫星互联网业务种类和需求的不断增加,提高卫星互联网的服务质量已引起人们的广泛关注。为了减少各种业务流量源汇聚对卫星通信系统的影响,自相似流量的预测变得十分重要。该文根据网络流量的自身特性,在传统神经网络预测模型的基础上,采用了一种更为切合其性质的长短期记忆网络预测模型对其进行流量预测。通过对ON/OFF物理模型产生流量进行训练,得到结果。通过图像、数据、误差指标多方面对比,进行预测结果评价。仿真结果表明:长短期记忆神经网络预测模型可以实现网络流量的预测,精度较高。

The Impact of Self-Similar Traffic on Network Delay

The effect of self-similar traffic on the delay of a single queue systemis studied through the use of the measured traffic and models as input process. Amodel-driven simulation-based method is then proposed for the computation of meanline delay in a network design. Both the hybrid-FGN and the FARIMA algorithmshave been used to synthesize self-similar sample paths. The comparison results withreal-traffic data sets firmly establish the usefulness of the proposed model-drivensimulation-based method. A practical database method is also introduced that helpsthe designer to determine the parameters in network design. This approach mayplay an important role in network design and analysis.

自相似数据流的Hurst指数小波求解法分析

该文针对具有自相似性的数据流的特征参数——Hurst指数的小波提取方法进行了系统的分析和研究,对现有提取Hurst指数的小波方法给予了归纳和总结。通过大量的实验和分析,研究了消失矩阶数、小波函数以及分解级数的选择对小波法提取Hurst指数的准确性的影响,并得出了一些有意义的结论。

基于EMD及ARMA的自相似网络流量预测

提出了一种基于ARMA(自回归滑动平均)模型的经验模式分解预测自相似网络流量的方法,进行了理论证明和仿真验证。结果表明,经验模式分解对长相关流量有去相关的作用,采用ARMA模型即可对自相似网络流量准确刻画,不但降低了算法的复杂度,而且预测精度高于径向基函数神经网络的预测精度。

自相似业务下光缓存设计

提出了在数据包长可变的自相似业务下研究光纤延时线所组成的光缓存问题的解析模型 .利用 Pareto到达过程来模拟自相似业务 .研究表明 ,缓存深度越大 ,包丢失率越低 ,而且存在与业务强度对应的最佳光纤延时线单位 ,使得在给定的缓存深度下 ,数据包丢失率最小 ,在业务强度和缓存深度相同时 ,Pareto到达比

基于模糊神经网络的网络业务量预测研究

利用神经网络(NN)的自学习能力以及模糊逻辑的动态性和及时性等特点,将模糊逻辑和 NN 有机地结合起来,构造出了五层模糊神经网络(FNN),并用训练 NN 的相应学习算法-BP 算法来训练网络。本文将 FNN 用于网络自相似业务预测研究中,并与单纯的 NN 算法相比较。仿真结果表明,FNN 能很好地预测复杂网络业务,与传统的 NN 算法相比,不仅收敛速度快,且得到更好的预测效果。本文为复杂网络业务流量预测研究提供了一种有效途径。

基于混沌预测的宽带DVB-RCS卫星接入信道动态分配方案研究

该文概述了基于卫星数字视频广播-反向信道系统(DVB-RCS)标准的宽带卫星通信网的特点,并重点研究了多址接入信道的动态分配策略,在利用混沌理论对网络中的自相似业务流量进行预测的基础上,提出了一种新型混合按需分配多址接入方案。该文还利用OPNET软件建立了仿真系统,用于比较不同方案的性能。仿真结果表明,在高信道负载以及网络中业务流具有较高自相似程度的条件下,该文所提出方案具有更优越的性能。

基于分形整形器的GPS系统性能确定上界研究

为了精确求解以自相似业务作为输入的网络性能确定边界问题,在GPS(generalized processor sharing)系统入口处利用分形整形器对自相似业务进行调控,提出了基于分形整形器的GPS系统模型,基于确定性网络演算理论推导了使用自相似业务作为输入时该GPS系统的缓存队列长度、延迟、延迟抖动以及有效带宽等性能确定上界模型。数值结果表明,基于分形整形器的GPS系统性能确定上界模型对自相似业务具有较好的性能。

自相似业务量下网络的排队性能分析

在研究通信网的网络性能、流量控制和资源配置时 ,业务源模型是其中一个十分重要的组成部分。近年来的研究发现数据网络中的业务量呈自相似特征 ,传统的能够精确描述电话网络的马尔可夫模型不再适用。业务量的自相似特征显著影响网络的流量控制与排队分析 ,已经引起人们的极大重视。采用G M 1排队模型对分组长度可变的网络的排队性能进行了分析和仿真 ,其中自相似业务量是通过叠加大量独立的到达间隔为Pareto重尾分布的ON OFF源来生成的。仿真结果表明 ,自相似业务量导致网络的排队性能劣化 ,这与有关文献对ATM交换的分析结果一致。