[Dataset] ilesansfil/wifidog (v. 2007-08-27)

top

the initial version

@MISC{ilesansfil-wifidog-2007-08-27,
  author = {Michael Lenczner and Benoit Grégoire and François Proulx},
  title = {{CRAWDAD} data set ilesansfil/wifidog (v. 2007-08-27)}, 
  howpublished = {Downloaded from http://crawdad.cs.dartmouth.edu/ilesansfil/wifidog},
  month = aug,  
  year = 2007
}
					

This data set contains user session traces which were collected from 
a large number of free Wi-Fi hotspots in Montréal, Québec, Canada  for three years.

Île Sans Fil (French for "Island Without Wires", also known as ISF), 
a non-profit organization, operates a network of free Wi-Fi hotspots in 
Montréal, Québec, Canada.  It now counts 140 hotspots and over 45,000 users.  
All of the hotspots are located in publicly accessible spaces. The ISF hospots
are deployed mostly in cafes, restaurants and bars, but also in libraries, 
funeral homes, doctors' offices, and BIAs (Business Improvement Agencies) 
to cover parks and sections of popular commercial streets.
For more information, please see http://www.ilesansfil.org

ISF started the WifiDog project to provide a solution for managing hotspots.
WifiDog is a software solution which is designed to meet the initial ISF 
requirements, such as content-centered nature of the auth server, 
the exceptionally extensive multilingual support, and the very thin gateway.
For more information, please http://dev.wifidog.org/

The ISF team collects user session data that are neccessary for the operation 
of some features (like location-aware content delivery, prior-view aware 
content delivery, viewing users online at the same hotspot, etc.) 
Other motivations include detecting and controlling bandwidth hogging, 
and knowing how, where, and how much their network is used.

WifiDog is composed of the following 2 components:
- the gateway is a daemon process that gets installed on every wireless router on a hotspot.
- the auth server is a web application that gets installed in a central location.

WifiDog enables a new hotspot site to configure the equipment by just plugging it 
in the existing network and entering the new router password. 

Existing networks, such as DSL and cable connections, have proven rather unreliable, 
so WifiDog makes monitoring easier in the following ways: 

- it graphically displays the status and location of each hotspot (see http://carte.ilesansfil.org/). 
- in case of fault, it sends mail to local support resources assigned to that hotspot.  
- it is completely unaffected by network address translation or firewalls upstream. 
- finally it does all this without any extra software (VPN, SNMP, etc).

The ISF team collects user session (i.e., between login and logout) data such as 
account (user) id, MAC address, login and logout time, hotspot id, and 
amount of data transferred incoming and outgoing. 

They also logs content display and click-through, but only for content being displayed 
on the auth server, and therefore the data are not included in this data set.

The ISF team anonymized account id (user_id), connection id (conn_id), user MAC address 
(user_mac), and hotspot id (node_id).  They performed hashing with a different random 
seed for every value. The code that generated the anonymised data can be found at

http://dev.wifidog.org/browser/trunk/wifidog-auth/wifidog/classes/StatisticReport/AnonymisedDataExport.php

[Traceset] ilesansfil/wifidog/session (v. 2007-08-27)

top

the initial version

@MISC{ilesansfil-wifidog-session-2007-08-27,
  author = {Michael Lenczner and Benoit Grégoire and François Proulx},
  title = {{CRAWDAD} trace set ilesansfil/wifidog/session (v. 2007-08-27)}, 
  howpublished = {Downloaded from http://crawdad.cs.dartmouth.edu/ilesansfil/wifidog/session},
  month = aug,  
  year = 2007
}
					

This traceset contains user sessions which were collected from 
a large number of free Wi-Fi hotspots in Montréal, Québec, Canada  for three years.

When a user connects to a gateway installed on a wireless router, the gateway 
opens a session on the auth server (that knows what hotspot the user is connecting
from), the auth server assigns a session number (connection id) and gives it 
to the gateway.  The gateway then counts the in and out traffic transferred during 
that session, and periodically sends it to the auth server for each connected user. 
Therefore, no merge of data from diffrent hotspots is necessary, the only place 
data is logged is on the auth server.

[Trace] ilesansfil/wifidog/session/04_07 (v. 2007-08-27)

top

the initial version

@MISC{ilesansfil-wifidog-session-04_07-2007-08-27,
  author = {Michael Lenczner and Benoit Grégoire and François Proulx},
  title = {{CRAWDAD} trace ilesansfil/wifidog/session/04_07 (v. 2007-08-27)}, 
  howpublished = {Downloaded from http://crawdad.cs.dartmouth.edu/ilesansfil/wifidog/session/04_07},
  month = aug,  
  year = 2007
}
					

User session traces which were collected from a large number of 
free Wi-Fi hotspots in Montréal, Québec, Canada  for three years.

The logged data are stored in a relational database in normalized form 
and retrieved in realtime as needed. There is no post-processing 
done on the data.

This trace contains 587782 user sessions for 69689 (distinct) users, 
which were collected from 206 hotspots for three years.

The trace includes postgresql compatible file of SQL statements.
The table created has the same column names, and the same datatype as 
the real data. The table schema is  as follows:

CREATE TABLE connections_anonymised	(
	conn_id text NOT NULL,
	timestamp_in timestamp,
	node_id text,
	timestamp_out timestamp,
	user_id text NOT NULL DEFAULT '',
	user_mac text,
	incoming int8,
	outgoing int8);

where 

	conn_id :  connection id
	timestamp_in : login time 
	node_id : hotspot id 
	timestamp_out : logout time 
	user_id : user account id 
	user_mac : user MAC address 
	incoming : incoming data transferred (Bytes) 
	outgoing : outgoing data transferred (Bytes)

[Author] Michael Lenczner

top

[Author] Benoit Grégoire

top

[Author] François Proulx

top