基于机器学习的内部延迟估计网络层析成像
发布时间:2024-07-03 01:45
随着互联网渗透到人们生活的方方面面(物联网),计算机网络变得日渐庞大、复杂。在这种情况下,用不影响监测网络性能的方式获得指标和度量值,并进行及时有效的网络监测和分析就变得至关重要。然而,测量网络中所有节点的网络流量是不切实际的,一种有前景的替代方案是仅在网络边缘进行测量,并从这些测量值中推断网络的内部行为。为了解决内部链路参数测量(例如时延和丢包率)的问题,本文采用网络层析(NT)技术,收集基于端到端测量的路径性能数据,然后使用统计计算的方法推断逻辑链路性能的概率分布。这种从端到端估计链路性能的技术既不需要内部网络的协作,也不依赖通信协议。此外,本文在网络层析成像技术上融合了一种新的统计方法,使得建模网络更容易估计内部链路性能参数的性能。这种方法就是机器学习(ML),尤其是线性回归模型。该技术能够在给定输入值(例如路径时延)后预测真实值(例如链路时延),比如说让模型从给定的样本数据(学习数据大约占总数据的80%)中学习,然后使用20%的数据(测试数据)验证模型。将估计得到的时延值与使用网络模拟器NS2对一个有线网络仿真生成的实际时延值进行比较,以两者的计算误差值(均方误差MSE)评估模...
【文章页数】:61 页
【学位级别】:硕士
【文章目录】:
摘要
ABSTRACT
LIST OF ABBREVIATIONS
Chapter 1 Introduction
1.1 Introduction
1.1.1 Network Tomography in the Internet
1.2 Problem Statement
1.3 Motivation
1.4 Contribution
1.5 Thesis Layout
Chapter 2 Literature Review
2.1 Introduction
2.1.1 Computer Network
2.1.2 Tomography
2.1.3 Network Tomograph
2.2 Delay Network Tomography
2.2.1 Estimation of Propagation Link Delay
2.2.2 Estimation of Link Delay Density
2.3 Traffic Network Tomography
2.3.1 Vardi's model
2.3.2 Vanderbei and lannone's model
2.3.3 Maximum likelihood parameter estimation
2.3.4 The EM algorithm
2.3.5 Other aspects for the Network Tomography
Chapter 3 Statistica Inference
3.1 Introduction
3.2 Categories of Machine Learning
3.2.1 Supervised Learning
3.2.2 Unsupervised Learning
3.2.3 Reinforcement Learning
3.3 Linear Regression
3.3.1 Simple Regression
3.3.2 Multiple Regression
3.4 The Loss Function
Chapter 4 Methodology
4.1 Introduction
4.2 Simulation methodology
4.2.1 Capabilities and limitations of ns2
4.2.2 Network simulation using ns2
4.2.3 Network Animator NAM
4.3 The Designed Model
4.3.1 Machine Learning implementation
4.4 End-To-end Delay
4.4.1 The measurement procedure
Chapter 5 Evaluation and Results
5.1 Introduction
5.2 Result and Discussion
Chapter 6 Conclusion and Future Work
6.1 Conclusion
6.2 Future Work
Bibliography
ACKNOWLEDGEMENT
PUBLICATIONS AND MASTER ACTIVITIES
本文编号:4000335
【文章页数】:61 页
【学位级别】:硕士
【文章目录】:
摘要
ABSTRACT
LIST OF ABBREVIATIONS
Chapter 1 Introduction
1.1 Introduction
1.1.1 Network Tomography in the Internet
1.2 Problem Statement
1.3 Motivation
1.4 Contribution
1.5 Thesis Layout
Chapter 2 Literature Review
2.1 Introduction
2.1.1 Computer Network
2.1.2 Tomography
2.1.3 Network Tomograph
2.2 Delay Network Tomography
2.2.1 Estimation of Propagation Link Delay
2.2.2 Estimation of Link Delay Density
2.3 Traffic Network Tomography
2.3.1 Vardi's model
2.3.2 Vanderbei and lannone's model
2.3.3 Maximum likelihood parameter estimation
2.3.4 The EM algorithm
2.3.5 Other aspects for the Network Tomography
Chapter 3 Statistica Inference
3.1 Introduction
3.2 Categories of Machine Learning
3.2.1 Supervised Learning
3.2.2 Unsupervised Learning
3.2.3 Reinforcement Learning
3.3 Linear Regression
3.3.1 Simple Regression
3.3.2 Multiple Regression
3.4 The Loss Function
Chapter 4 Methodology
4.1 Introduction
4.2 Simulation methodology
4.2.1 Capabilities and limitations of ns2
4.2.2 Network simulation using ns2
4.2.3 Network Animator NAM
4.3 The Designed Model
4.3.1 Machine Learning implementation
4.4 End-To-end Delay
4.4.1 The measurement procedure
Chapter 5 Evaluation and Results
5.1 Introduction
5.2 Result and Discussion
Chapter 6 Conclusion and Future Work
6.1 Conclusion
6.2 Future Work
Bibliography
ACKNOWLEDGEMENT
PUBLICATIONS AND MASTER ACTIVITIES
本文编号:4000335
本文链接:https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/4000335.html