MLInNet/index.html at master · reinnet/MLInNet · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
<!doctype html>
<html>
  <head>
    <meta charset="utf-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0, maximum-scale=1.0, user-scalable=no">

    <title>MLInNet</title>

    <link rel="icon" type="image/icon" href="img/favicon.png">

    <link rel="stylesheet" href="css/share.css">
    <link rel="stylesheet" href="css/reveal.css">
    <link rel="stylesheet" href="css/theme/material.css">

    <!-- Theme used for syntax highlighting of code -->
    <link rel="stylesheet" href="lib/css/zenburn.css">

    <!-- Printing and PDF exports -->
    <script>
      var link = document.createElement( 'link' );
      link.rel = 'stylesheet';
      link.type = 'text/css';
      link.href = window.location.search.match( /print-pdf/gi ) ? 'css/print/pdf.css' : 'css/print/paper.css';
      document.getElementsByTagName( 'head' )[0].appendChild( link );
    </script>
  </head>
  <body>
    <div class="reveal">
      <div class="slides">
        <section data-background-image="img/background.jpeg">
          <h1 style="color: white">Machine Learning in Networking</h1>
          <h3 style="color: white">Parham Alvani</h3>
          <h5><a href="https://1995parham.github.io" >@1995parham</a></h5>
        </section>
        <section>
          <div class="toc"></div>
        </section>
        <section>
          <div class="toc" data-selected="0"></div>
        </section>
        <section>
          <h3>Introduction</h3>
          <ul>
            <li>Networks needs more intelligence to deploy</li>
            <li>Traditional networks are <span class="material-italic">distributed</span></li>
            <li>Each node, such as router or switch, can only view and act over a <span class="material-italic">small partion</span> of the system.</li>
            <li>Learning from the nodes that have only a small partial view of the complete system to perform contol beyond the local domain is <span class="material-bold">very complex</span></li>
          </ul>
        </section>
        <section>
          <h3>Introduction</h3>
          <ul>
            <li>SDN decouples the control plane and the data plane</li>
            <li>The network resources in SDN are managed by a logically centralized controller</li>
          </ul>
        </section>
        <section>
          <h3>Software Defined Networking</h3>
          <img src="img/sdn.jpg" alt="sdn">
        </section>
        <section>
          <h3>Machine Learning in SDN</h3>
          <ul>
            <li>Traffic Classification</li>
            <li>Routing Optimization</li>
            <li>QoS/QoE Prediction</li>
            <li>Resource Management</li>
            <li>Security</li>
          </ul>
        </section>
        <section>
          <div class="toc" data-selected="1"></div>
        </section>
        <section>
          <ul>
            <li>Traffic classificiation is an important network function, which provides a way to perform fine-grained network management by identifying different flow types.</li>
            <li>In general <span class="material-bold">supervised and semi-supervised</span> learning algorithms can be used</li>
            <li>Deep Packet Inspection (DPI) is a common method to label traffic flows, but it incurs high computational cost when a large number of traffic flows are labeled</li>
          </ul>
        </section>
        <section>
          <div class="toc" data-selected="2"></div>
        </section>
        <section>
          <ul>
            <li>Routing is a fundamental network function</li>
            <li>Supervised Learning for obtaining the optimal huristic-like routing solution</li>
            <li>RL algorithms without labeled taining datasets can have flexible optimization targets</li>
            <li>The state space is composed of <span class="material-bold">network and traffic states</span></li>
          </ul>
        </section>
        <section>
          <div class="toc" data-selected="3"></div>
        </section>
        <section>
          <ul>
            <li>QoS prediction aims to discover the quantitative correlations between KPIs and QoS parameters</li>
            <li>QoS parameters are generally continuous data (Regression)</li>
            <li>As an example here we can use NN-Model for network delay estimation instead of M/M/1 model</li>
            <li>QoE prediction aims to discover the quantitative correlations between QoS parameters and QoE values</li>
            <li>QoE values are generally discrete data (Classification)</li>
          </ul>
        </section>
        <section>
          <div class="toc" data-selected="4"></div>
        </section>
        <section>
          <ul>
            <li>Efficient network resource management is the primary requirement of network operators to improve network performance</li>
            <li>The <span class="material-italic">data plane</span> resource allocation problem is generally considered as a decision-making task.</li>
            <li><span class="material-bold">RL</span> and <span class="material-bold">ML-based game theory</span> are two effective approaches</li>
            <li>RL in single-tenancy</li>
            <li>ML-based game theory in multi-tenancy</li>
            <li>The mapping between the resource consumption and control message rate is very important for <span class="material-italic">control plane</span> resource allocation</li>
          </ul>
        </section>
        <section>
          <div class="toc" data-selected="5"></div>
        </section>
        <section>
          <ul>
            <li>An <span class="material-bold">Instrusion Detection System (IDS)</span> is a device or software application and its objective is to monitor the events in a network system and identify possible attacks</li>
            <li>signature-based IDS and anomaly-based IDS</li>
            <li>Machine learning methods are widely used in anomaly-based IDS by trainning a model to identify normal activities and intrusions</li>
            <li>Intrusion detection problem can be considered as a classification task</li>
          </ul>
        </section>
        <section>
          <div class="toc" data-selected="6"></div>
        </section>
        <section>
          <ul>
            <li>High-quality Training Datasets</li>
            <li>Distributed Multi-controller Platform</li>
            <li>Improving Network Security</li>
            <li>Cross-layer Network Optimization</li>
            <li>Incrementally Deployed SDN</li>
          </ul>
        </section>
        <section data-background-image="img/thank-you.jpeg">
        </section>
      </div>
    </div>
    <script src="lib/js/head.min.js"></script>
    <script src="js/reveal.js"></script>

    <script>
      // More info https://github.com/hakimel/reveal.js#configuration
      Reveal.initialize({
        history: true,
        controls: true,
        slideNumber: true,

        margin: 0.3,

        // More info https://github.com/hakimel/reveal.js#dependencies
        dependencies: [
          {
            src: 'plugin/highlight/highlight.js',
            async: true,
            condition: function() {
              if (typeof hljs !== undefined) {
                return true;
              } else {
                return false;
              }
            },
            callback: function() { hljs.initHighlightingOnLoad(); }
          },
          {
            src: 'plugin/toc.js',
            async: true,
            condition: function() {
              if (typeof toc !== undefined) {
                return true;
              } else {
                return false;
              }
            },
            callback: function() { toc.initToCOnLoad([
              "Introduction",
              "Traffic Classification",
              "Routing Optimization",
              "QoS/QoE Prediction",
              "Resource Management",
              "Security",
              "Challenges"
            ]); }
          }
        ]
      });
    </script>

    <!-- logos and advertising -->
    <a href="https://github.com/Bambil/MLInNet" target="_blank"><img style="position: absolute; top: 0; right: 0; border: 0;" src="img/forkme_right_orange.png" alt="Fork me on GitHub"></a>
  </body>
</html>