Artificial Intelligence and Deep Learning for Computer Network: Management and Analysis aims to systematically collect quality research spanning AI, ML, and Deep Learning (DL) applications to diverse sub-topics of computer networks, communications, and security, under a single cover. It also aspires to provide more insights on the applicability of the theoretical similitudes, otherwise a rarity in many such books.
In recent years, particularly with the advent of Deep Learning (DL), new avenues have opened up to handle today’s most complex and very dynamic computer networks, and the large amount of data (often real time) that they generate. Artificial Intelligence (AI) and Machine Learning (ML) techniques have already shown their effectiveness in different networks and service management problems, including, but not limited to, cloud, traffic management, cybersecurity, etc. There exist numerous research articles in this domain, but a comprehensive and self-sufficient book capturing the current state-of-the-art has been lacking. The book aims to systematically collect quality research spanning AI, ML, and Deep Learning (DL) applications to diverse sub-topics of computer networks, communications, and security, under a single cover. It also aspires to provide more insights on the applicability of the theoretical similitudes, otherwise a rarity in many such books.
In the first chapter, application of Machine Learning to traffic management, in particular classification of domain name service (DNS) query packets over a secure (encrypted) connection is proposed. This important problem is challenging to solve because the relevant fields in the packet header and body that allows easy classification are not available in plaintext in an encrypted packet. An accurate Deep Learning model and a support vector machine (SVM)–based ML model is based on well-chosen features that were constructed to solve this problem.
Features:
A diverse collection of important and cutting-edge topics covered in a single volume.
Several chapters on cybersecurity, an extremely active research area.
Recent research results from leading researchers and some pointers to future advancements in methodology.
Detailed experimental results obtained from standard data sets.
This book serves as a valuable reference book for students, researchers, and practitioners who wish to study and get acquainted with the application of cutting-edge AI, ML, and DL techniques to network management and cyber security.
Author(s): Sangita Roy, Rajat Subhra Chakraborty, Jimson Mathew
Series: Chapman & Hall/Distributed Computing and Intelligent Data Analytics Series
Publisher: CRC Press
Year: 2023
Language: English
Pages: 137
Cover
Half Title
Series Page
Title Page
Copyright Page
Contents
Preface
About the Editors
Contributors
1. Deep Learning in Traffic Management: Deep Traffic Analysis of Secure DNS
1.1 Introduction
1.2 Survey on DoH, DoT, and Machine Learning Classification
1.3 Implementation (Diff Models and All)
1.3.1 Dataset
1.3.1.1 DoH vs. Non-DoH Dataset
1.3.1.2 Malicious vs. Non-Malicious DoH
1.3.2 Feature Engineering
1.3.3 Classification Models
1.3.3.1 The Keras Sequential Model
1.3.3.2 The SVM Model
1.4 Results and Analysis (with Graphs)
1.5 Conclusion
References
2. Machine Learning-Based Approach for Detecting Beacon Forgeries in Wi-Fi Networks
2.1 Introduction
2.2 Problem Statement
2.3 Related Work
2.4 Brief Introduction to the Models
2.4.1 SVM
2.4.2 k-NN
2.4.3 Random Forest
2.4.4 Multilayer Perceptron (MLP)
2.4.5 CNN
2.5 Dataset Generation
2.5.1 Beacon Forgery
2.5.2 Beacon Flooding
2.5.3 De-authentication Attack
2.5.4 Attack Modeling
2.5.4.1 Feature Extraction
2.6 Dataset Classification
2.7 Evaluation
2.7.1 Analyzis of Results
2.8 Conclusion and Future Work
References
3. Reinforcement Learning-Based Approach Towards Switch Migration for Load-Balancing in SDN
3.1 Introduction
3.2 Literature Survey
3.3 Load Balancing in SDN
3.3.1 Problem Formulation
3.4 Knowledge-Defined Networking
3.4.1 Classical SDN Architecture
3.4.2 SDN Architecture with Knowledge Plane
3.5 Load Balancing Using Reinforcement Learning
3.5.1 Reinforcement Learning
3.5.2 Markov Decision Process (MDP)
3.5.3 Problem Formulation Using Markov Decision Process
3.5.3.1 State Space (S)
3.5.3.2 Action Space (A)
3.5.3.3 Reward Function (R)
3.5.4 Q-Learning
3.5.4.1 Exploration and Exploitation Trade-off
3.6 Methodology
3.6.1 Random Approach
3.6.2 Q-Learning with ϵ-greedy
3.7 Results and Implementation
3.7.1 Experiment Setup
3.7.2 Evaluation Metric
3.7.3 Experimental Results
3.8 Conclusion and Future Work
References
4. Green Corridor over a Narrow Lane: Supporting High-Priority Message Delivery through NB-IoT
4.1 Introduction
4.1.1 Challenges in Delay-Sensitive Traffic Scheduling over NB-IoT
4.1.2 Contribution of This Work
4.1.3 Organization of the Paper
4.2 Related Works
4.3 NB-PTS: System Model and Design Details
4.3.1 Queueing Model Description
4.3.1.1 Solution Approach in NB-PTS
4.3.1.2 Derivation of Target Mean Delay
4.3.2 Estimation of Queue Threshold Value
4.3.2.1 ϵ-greedy Policy
4.3.3 Calculation of Target Mean Delay
4.3.4 Metric for Scheduling
4.3.5 Knowledge Base
4.3.6 Scheduling in NB-PTS
4.3.7 Time-Bound Analysis
4.4 Performance Analysis
4.4.1 Baseline Mechanisms
4.4.2 Prioritized Traffic Generation
4.4.3 Implementation Details
4.4.4 Analysis of Throughput
4.4.4.1 Overall Average Throughput
4.4.5 Analysis of Packet Loss Rate and Packet Delay
4.4.5.1 PLR and Delay of Individual Priority-Based Traffic
4.4.6 Convergence Behavior
4.4.7 Average Delay Distribution
4.4.8 Impact of Number of UEs on Consumed Subframes
4.4.9 Impact of Number of UEs on Computational Time
4.5 Conclusion
References
5. Vulnerabilities Detection in Cybersecurity Using Deep Learning-Based Information Security and Event Management
5.1 Introduction
5.2 Literature Survey
5.2.1 Research Gap
5.2.2 Importance of the Chapter in the Context of Current Status
5.3 Pictorial Representation of Cybersecurity Working Model
5.3.1 The Proposed Approach
5.3.2 Apply LSTM
5.3.3 Algorithm for LSTM
5.3.4 LSTM Implementation for Vulnerabilities Detection
5.3.5 Results
5.4 Conclusion
References
6. Detection and Localization of Double-Compressed Forged Regions in JPEG Images Using DCT Coefficients and Deep Learning-Based CNN
6.1 Introduction
6.1.1 Motivation and Objectives
6.1.2 Our Contributions
6.2 Related Background
6.2.1 Overview of JPEG Image Compression
6.2.2 JPEG Attack Model
6.2.3 Related Works
6.3 Deep Learning-Based Forensic Framework for JPEG Double-Compression Detection
6.3.1 JPEG DCT Coefficients Extraction and Selection
6.3.2 CNN Architecture
6.4 Localizing Double JPEG Compressed Forged Regions
6.4.1 JPEG Double-Compression Region Localization
6.4.2 Experimental Results for JPEG Double-Compression Localization
6.5 Conclusion
References
Index
