main.tex
7.58 KB
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
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
\documentclass[a0paper]{tikzposter}
\usepackage{url}
\usepackage[hidelinks]{hyperref}
\usepackage{booktabs}
\usepackage{subcaption}
\usepackage[usenames,dvipsnames]{color}
\usepackage{fontspec}
\usepackage{pgf}
\usepackage[absolute, verbose, overlay]{textpos}
\usepackage{caption}
\usepackage{array}
\usepackage{fancybox}
\usepackage{graphicx}
\usepackage{biblatex}
\usepackage{enumerate}
\usepackage{ulem}
\usepackage[absolute,overlay]{textpos}
\graphicspath{{figs/}}
\bibliography{main}
\captionsetup{labelformat=empty}
\setlength{\TPHorizModule}{1cm}
\setlength{\TPVertModule}{1cm}
\usetheme{Default}
%\usetheme{Rays}
%\usetheme{Basic}
%\usetheme{Simple}
%\usetheme{Envelope}
%\usetheme{Wave}
%\usetheme{Board}
%\usetheme{Autumn}
%\usetheme{Desert}
%\usecolorpalette{BrownBlueOrange}
%\usecolorstyle{Germany}
%\usecolorstyle{Denmark}
%\usecolorstyle{Russia}
%\usecolorstyle{Default}
%\usecolorstyle{Australia}
%\usecolorstyle{Britain}
%\usecolorstyle{Sweden}
%\usecolorstyle{Spain}
\newcommand{\PS}{\textsc{PocketSniffer}}
\title{\parbox{\linewidth}{\centering Crowdsourcing Access Network Spectrum Allocation Using Smartphones}}
\author{%
Jinghao Shi$^\dag$, Zhangyu Guan$^\dag$$^\S$, Chunming Qiao$^\dag$, Tommaso
Melodia$^\S$\\Dimitrios Koutsonikolas$^\dag$ and Geoffrey Challen$^\dag$\\
\vspace*{5mm}
\url{pocketsniffer@blue.cse.buffalo.edu}
}
\institute{%
$^\dag$University at Buffalo\quad$^\S$Northeastern University
}
%\titlegraphic{\includegraphics[width=0.6\textwidth]{ub_logo}}
\begin{document}
\maketitle
\begin{columns}
\column{0.5}
\block{1. So Many APs, So Few Channels}{%
\begin{tikzfigure}[Crowded RF space.]
\begin{subfigure}{0.45\colwidth}
\includegraphics[width=\textwidth]{channel}
\end{subfigure}\hspace{0.01\colwidth}%
\begin{subfigure}{0.45\colwidth}
\includegraphics[width=\textwidth]{crowd}
\end{subfigure}
\end{tikzfigure}
\vspace{5mm}
More and more wireless devices are competing for the limited spectrum
resources. Recent studies have shown the value of client-side measurements
in determining better channel assignment. Yet \textbf{how to collect data
from clients without interrupting their normal usage} remains to be an
open question. To address this problem, we propose \PS{}---\textbf{a
framework for client-side measurements collection}. The key idea is to
\textbf{use inactive smartphones to perform measurements on behalf of
nearby devices} to improve spectrum allocation. \PS{} also collects the
network measurements naturally generated by smartphones to monitor the
health and performance of large-scale wireless networks.
}
\block{2. Why Smartphones?}{%
\begin{tikzfigure}[Smartphones are carried with you, always on but mostly
idle.]
\begin{subfigure}[t]{0.3\colwidth}
\centering
\includegraphics[width=\textwidth]{carried}
\end{subfigure}\hspace{0.01\colwidth}%
\begin{subfigure}[t]{0.28\colwidth}
\centering
\includegraphics[width=\textwidth]{always_on}
\caption{\large\textbf{Always on.}}
\end{subfigure}\hspace{0.01\colwidth}%
\begin{subfigure}[t]{0.27\colwidth}
\centering
\includegraphics[width=\textwidth]{mostly_idle}
\caption{\large\textbf{Mostly idle.}}
\end{subfigure}
\end{tikzfigure}
Smartphones are carried with people, which enables them to measure the
network conditions from \textbf{actual user's point of view}. Smartphones are
\textit{always on} yet \textit{mostly idle}, making it possible to \textbf{conduct
measurements without interrupting users' normal usage}.
}
\block{3. System Design}{%
\begin{tikzfigure}[\PS{} System]
\includegraphics[width=0.6\colwidth]{system}
\end{tikzfigure}
}
\block{4. Challenges and Approaches}{%
\begin{itemize}
\item Proximity detection---use measurements from one device to
approximate the another.
\begin{itemize}
\item Explore proximity in both physical and Wifi signature space.
\end{itemize}
\item Measurement validation---defeat false measurements.
\begin{itemize}
\item Leverage trusted APs.
\end{itemize}
\item Incentives---encourage measurements.
\begin{itemize}
\item Bob's phone help Bob's laptop.
\item Measurements for QoS.
\end{itemize}
\end{itemize}
}
\column{0.5}
\block{5. Coordination Scenarios}{%
\begin{tikzfigure}
\centering
\begin{subfigure}{0.3\colwidth}
\centering
\includegraphics[width=\textwidth]{single_sp_managed}
\captionof{figure}{Single SP, managed clients\\(e.g., home Wifi network).}
\end{subfigure}\hspace*{0.1\colwidth}%
\begin{subfigure}{0.3\colwidth}
\centering
\includegraphics[width=\textwidth]{single_sp_unmanaged}
\captionof{figure}{Single SP, unmanaged clients\\(e.g., campus Wifi
network).}
\end{subfigure}
\end{tikzfigure}
\begin{tikzfigure}
\begin{subfigure}{0.4\colwidth}
\centering
\includegraphics[width=\textwidth]{multiple_sp_managed}
\captionof{figure}{Multiple SP, managed clients\\(e.g., overlapping home
network).}
\end{subfigure}\hspace*{0.01\colwidth}%
\begin{subfigure}{0.4\colwidth}
\centering
\includegraphics[width=\textwidth]{multiple_sp_unmanaged}
\captionof{figure}{Multiple SP, unmanaged clients\\(e.g., overlapping
enterprise network).}
\end{subfigure}
\end{tikzfigure}
}
\block{6. Discarded Treasure}{%
\begin{tikzfigure}
\centering
\begin{minipage}[T]{0.4\colwidth}
\vspace{-3cm}
To cope with rapid user mobility, smartphones already perform aggressive
network exploration. For example, the right figure shows that Android
will scan every 15~s or 60~s depending on the device's Wifi connection.
This is a treasure that is being discarded. \PS{} passively collects
such measurements in an energy neutral manner, to help monitor large
scale wireless networks.
\end{minipage}\hspace{0.03\colwidth}%
\begin{minipage}[T]{0.5\colwidth}
\includegraphics[width=\textwidth]{scan_interval}
\label{fig:scan}
\end{minipage}
\end{tikzfigure}
}
\block{7. Prototype}{%
\begin{tikzfigure}
\centering
\begin{subfigure}{0.42\colwidth}
\centering
\includegraphics[width=\textwidth]{channel_switch}
\end{subfigure}\hspace{0.03\colwidth}%
\begin{subfigure}{0.42\colwidth}
\centering
\includegraphics[width=\textwidth]{switch.pdf}
\end{subfigure}
\end{tikzfigure}
We hacked the driver of Nexus 5 smartphones to support Wifi monitor
mode. We also set up several TP-LINK WDR3500 wireless routers running
OpenWRT as \PS{} APs. We conducted a simple channel adaptation experiment as
follows:
\begin{enumerate}
\item \textbf{0--7~s}\quad TCP link between \PS{} AP and device $A$.
\item \textbf{7--75~s}\quad $B$ jams the channel, causing bandwidth
degradation and jitter of $A$.
\item \textbf{75--100~s}\quad AP switches to a less congested channel with
the help of nearby device $C$.
\end{enumerate}
Throughout the experiment, \textbf{device $A$'s active session is not
affected---neither $A$'s association with AP or TCP link is interrupted by
the channel switch}.
This experiment demonstrate the feasibility of \PS{} in single SP, managed
clients scenario. We're extending the system to other three scenarios.
}
\end{columns}
\end{document}