Intro.

Geoffrey Challen
1 parent 20901b6d
Showing 3 changed files with 74 additions and 78 deletions
abstract.tex
introduction.tex
investigation.tex
@@ -20,12 +20,12 @@
 existing human relationships and can be maintained without elaborate
 reputation mechanisms.
- To evaluate the effectiveness of reciprocal \wifi{} sharing, we analyze
- 21~M \wifi{} scans collected from 254~smartphones over 5~months. Our
- results show that even in a sparsely-populated suburban area, reciprocal
- \wifi{} sharing is needed. And surprisingly, several reciprocal \wifi{}
- sharing opportunities exist even among our extremely-small sample of users.
- Motivated by these results, we present the design of \wisefi{}, a system
- enabling reciprocal \wifi{} sharing.
+ To evaluate the potential for reciprocal \wifi{} sharing, we analyze 21~M
+ \wifi{} scans collected from 254~smartphones over 5~months. Our results
+ show that even in a sparsely-populated suburban area, reciprocal \wifi{}
+ sharing can be beneficial. And surprisingly, we detected several reciprocal
+ \wifi{} sharing opportunities even within our tiny user sample. Motivated
+ by these results, we present the design of \wisefi{}, a system enabling
+ reciprocal \wifi{} sharing.
 \end{abstract}
@@ -14,7 +14,10 @@ other nearby private home APs.
 \begin{figure}[t]
 %
- \includegraphics[width=\columnwidth]{./figures/motivation.pdf}
+ \centering
+ \includegraphics[width=0.9\columnwidth]{./figures/motivation.pdf}
+ %
+ %\vspace*{-0.1in}
 %
 \caption{\textbf{Example of Reciprocal \wifi{} Sharing.} Solid arrows
 represent weak connections, while dashed lines represent strong
@@ -22,60 +25,54 @@ other nearby private home APs.
 %
 \label{fig:motivation}
 %
+ \vspace*{-0.1in}
 \end{figure}
 Unfortunately, uncoordinated deployment of overlapping private networks can
 create interference that degrades performance, which may then cause users to
-respond in ways that further exacerbate the problem. Consider Alice and Bob's
-neighboring apartments shown in Figure~\ref{fig:motivation}. Alice has
-deployed her AP in her living room, while Bob has deployed his in his
-bedroom. Due to the proximity of their apartments, Alice receives a stronger
-signal from Bob's router when she is in her bedroom. But because she cannot
-connect to Bob's router, she must either use the lower-bandwidth connection
-to her existing AP or deploy an additional AP in her bedroom. Both of these
-options generate additional wireless interference for her neighbors,
-including Bob. And while we have used Alice as an example, Bob also faces the
-same poor choice.
-
-However, due to factors such as blockage or fading in wireless signal
-propagation, home \wifi{} AP usually does not provide equally satisfying \wifi{}
-coverage at all places within the house. Instead, it is likely that the user
-receives better \wifi{} signal from a neighbor's AP at certain spots. For
-instance, consider Alice and Bob who live in neighbor apartments, as shown in
-Figure~\ref{fig:motivation}, each of them receives a stronger \wifi{} signal
-from the other's home AP than their own at certain places within their
-apartments, revealing a \textit{reciprocal} \wifi{} sharing opportunity where
-both parties can improve their \wifi{} performance by allowing each other to
-access their own private networks.
-
-Compared to traditional community networks such as FON~\cite{fon} or
-OpenWireless~\cite{openwireless}, such reciprocal sharing opportunity has
-several unique properties that make it interesting to explore. First, such
-opportunity is usually \textit{immediate} between two physically colocated
-parties, such as two neighbors. This helps relief the concerns of sharing
-network to random strangers in traditional community networks and makes it more
-practical to establish the sharing. Second, bonding to physical colocation
-relationship makes the opportunity \textit{stable} over time, enabling
-asynchronous fair sharing over longer period of time.
+respond in ways that further exacerbate the problem. Consider Alice's/Bob's
+apartment shown in Figure~\ref{fig:motivation}. Alice/Bob has deployed
+her/his AP in her/his living room/bedroom. Due to the proximity of their
+apartments, Alice/Bob receives a stronger signal from Bob's/Alice's router
+when she/he is in her/his bedroom/living room. But because Alice/Bob cannot
+connect to Bob's/Alice's router, she/he must either use the lower-bandwidth
+connection to her/his existing AP or deploy an additional AP in her/his
+bedroom/living room. Both options generate additional wireless interference
+for her/his neighbors, including Bob/Alice.
+Ideally, Alice/Bob would allow Bob/Alice to use her/his router. Obviously
+this solution requires less hardware. But it also improves performance while
+reducing interference and client energy consumption, both by allowing the APs
+to coordinate overlapping transmissions and by allowing clients to achieve
+higher bitrates at lower transmission powers. We refer to this
+mutually-beneficial arrangement as \textit{reciprocal \wifi{} sharing}.
-Nevertheless, there are several challenges in fulfilling the vision of
-reciprocal \wifi{} sharing shown in Figure~\ref{fig:motivation}. First, although
-the motivating example is inspired by the authors' own experience, it is not
-clear how often such opportunity exists for broader range of users in real life
-scenarios. Second, suppose the sharing opportunity does exist and is detected,
-there is no systematic solutions to enable the \wifi{} sharing without
-compromising the security and privacy of user's home network. Finally, after the
-\wifi{} sharing is established, it is challenging to ensure that the
-relationship remains reciprocal for both parties.
+Reciprocal \wifi{} sharing has benefits compared to attempts to use private
+APs to establish community networks such as FON~\cite{fon} or
+OpenWireless~\cite{openwireless}. Reciprocal \wifi{} sharing opportunities
+are more likely to align with existing human relationships, such as this
+example involving two neighbors, rather than requiring users to open their
+private networks to strangers. And because reciprocal \wifi{} sharing
+involves only pairwise cooperation, agreements can be established and
+monitored without the elaborate reputation systems or credit mechanisms
+required to prevent freeloading in large communities. Once Alice notices that
+the sharing agreement with Bob is no longer beneficial---either because she
+no longer needs his connection or because he is degrading her service to the
+point where it is no longer useful---she can immediately terminate it.
-To address these challenges, we first present extensive analysis of the
-\PhoneLab{} \wifi{} dataset which contains \num{21192417} scan results from 254
-smartphones over 5 months (Section~\ref{sec:investigation}). The results show
-that such reciprocal \wifi{} sharing opportunities does exists even in a spatial
-sparse dataset. Inspired by the results, we present the design of \wisefi{}
-(Section~\ref{sec:design}), a system that detects such reciprocal \wifi{}
-sharing opportunities using smartphones, enables \wifi{} sharing on APs with or
-without guest network support, and ensures the sharing remain reciprocal.
-Finally, we discuss some open challenges in implementing such a system and point
-directions for future works (Section~\ref{sec:challenges}).
+But how often is reciprocal \wifi{} sharing beneficial and possible in
+practice? To explore these questions, we begin in
+Section~\ref{sec:investigation} by analyzing a dataset collected on the
+\PhoneLab{}~smartphone testbed containing \num{21192417} \wifi{} scan results
+from 254~smartphones over 5~months (\S~\ref{sec:investigation}). Despite the
+fact that the geographic extent of the dataset is suburban Buffalo, which as
+a city has a population density an order of magnitude lower than
+densely-populated areas like Manhattan, we still find that many users would
+benefit from being able to connect to neighboring private networks. Even more
+surprisingly, despite monitoring only several hundred users we were still
+able to identify several reciprocal \wifi{} sharing opportunities in our tiny
+sample. Motivated by these results Section~\ref{sec:design} presents the
+design of \wisefi{}, a system addressing the practical challenges of
+establishing and monitoring reciprocal \wifi{} sharing agreements. We
+conclude by identifying some open challenges in implementing such a system as
+future work in Section~\ref{sec:challenges}.
@@ -2,12 +2,12 @@
 \label{sec:investigation}
 To investigate reciprocal sharing opportunity in real life scenarios, we
-obtained a \wifi{} scan result dataset from \PhoneLab{}
-(\S\ref{subsec:phonelab}). We first discuss some heuristics to identify the home
-AP for each device (\S\ref{subsec:homeap}). Then we show the RSSI comparison
-between a user's home and neighbor APs (\S\ref{subsec:better}). Finally, we
-explore the reciprocal sharing relationships in the dataset
-(\S\ref{subsec:reciprocal}).
+obtained a \wifi{} scan result dataset from
+\PhoneLab{}\footnote{\url{http://www.phone-lab.org}} (\S\ref{subsec:phonelab}).
+We first discuss some heuristics to identify the home AP for each device
+(\S\ref{subsec:homeap}). Then we show the RSSI comparison between a user's
+home and neighbor APs (\S\ref{subsec:better}). Finally, we explore the
+reciprocal sharing relationships in the dataset (\S\ref{subsec:reciprocal}).
 \subsection{PhoneLab \wifi{} Dataset}
 \label{subsec:phonelab}
@@ -31,20 +31,19 @@ explore the reciprocal sharing relationships in the dataset
 \label{tab:summary}
 \end{table}
-\PhoneLab{}\cite{phonelab-sensemine13} is a public smartphone platform testbed
-operated at the University at Buffalo. Several hundreds of participants carry
-instrumented Nexus 5 smartphones as their primary device. In particular, the
-smartphone platform was modified to log each \wifi{} scan result and \wifi{}
-connection events naturally generated by the Android system. Note that from data
-collection point of view, platform instrumentation is not necessary, and the
-same information can also be logged by applications with appropriate
-permissions. Table~\ref{tab:summary} summarizes the \PhoneLab{} \wifi{} dataset.
-
-A \wifi{} scan result represents the device's network visibility, and consists
-of multiple entries---each corresponds to one \wifi{} AP the device observed.
-The content of one entry includes: (1) beacon timestamp, (2) AP SSID and BSSID,
-(3) AP channel and (4) RSSI. Additionally, the timestamp when the scan was
-performed is also logged.
+\PhoneLab{}\cite{phonelab-sensemine13} is a public smartphone platform
+testbed operated at the University at Buffalo. Several hundreds of
+participants carry instrumented Nexus 5 smartphones as their primary device.
+In particular, the smartphone platform was modified to log each \wifi{} scan
+result and \wifi{} connection events naturally generated by the Android
+system. Note that from data collection point of view, platform
+instrumentation is not necessary, and the same information can also be logged
+by applications with appropriate permissions. A \wifi{} scan result
+represents the device's network visibility, and consists of multiple
+entries---each corresponds to one \wifi{} AP the device observed. The content
+of one entry includes: (1) beacon timestamp, (2) AP SSID and BSSID, (3) AP
+channel and (4) RSSI. The timestamp when the scan was performed is also
+logged. Table~\ref{tab:summary} summarizes the \PhoneLab{} \wifi{} dataset.
 \subsection{Home AP Detection}
 \label{subsec:homeap}