User assisted analysis of cellular network structures 26.10.2011 - - PowerPoint PPT Presentation

Network Architectures and Services, Georg Carle Faculty of Informatics Technische Universität München, Germany

User assisted analysis of cellular network structures

26.10.2011

Author: Simon Wagner Advisor: Johann Schlamp Supervisor: Prof. Georg Carle

Simon Wagner - User assisted analysis of cellular network structures 2

What did we do?

• We developed an Android application to collect

information about the user's network

Simon Wagner - User assisted analysis of cellular network structures 3

Why did we do it?

Value for the user:

• User can measure his mobile connection
• Collected data might be interesting for the user

himself (e.g. comparing network availability for a provider) Value for us:

• We get to know how users access the internet
• n mobile devices
• We can take a look how users are using their

mobile phones (e.g. GPS availability and accuracy)

Simon Wagner - User assisted analysis of cellular network structures 4

How did we do it?

• Measurements are conducted in the

background without the users interactions. They are started by an Android background service that is started in a certain interval

• The user can deactivate the measurements if

he wants to

• The user can measure his down- and upstream
• n his request
• The measured speed will then be reported to

the user and us

Simon Wagner - User assisted analysis of cellular network structures 5

How did we do it?

• Measurements will get uploaded as JSON to
• ur servers as soon as an internet connection is

available and will be stored in a CouchDB instance

• Therefore new information can be collected

without changing the storage backend

• CouchDB provides a safe storage

Collector store upload

Simon Wagner - User assisted analysis of cellular network structures 6

How did we do it?

• Another server provides a web interface to our data
• Interested users can take a look at our collected data
• Presents some interesting data like recorded speed, availability
• f different network types, etc.
• Synchronizes from time to time with the storage server and

updates its own database

Postgres DB

synchronize

Simon Wagner - User assisted analysis of cellular network structures 7

Web interface - Statistics

The webinterface provides information about the performance of different networks and also serves informations about the collected statistics

Simon Wagner - User assisted analysis of cellular network structures 8

Web interface - Map

The location of our measurement points is shown together with detected basestations (blue) and basestations from OpenCellID (white)

Simon Wagner - User assisted analysis of cellular network structures 9

User distribution

• Users are mainly from the USA
• Germany ranks second
• Rest comes from Europe

Simon Wagner - User assisted analysis of cellular network structures 10

Number of active users

Overall number of users: 459 Collected measurements: 242299

(as of September 1st)

Number of active users

Simon Wagner - User assisted analysis of cellular network structures 11

User distribution

The distribution of Android versions does not differ much from the official data presented below Measured Android OS distribution amongst our users:

Simon Wagner - User assisted analysis of cellular network structures 12

Geolocation services

• When developing a location-aware Android application, you should

use both geolocation services

• GPS has a reliable accuracy, network based location's accuracy is

fluctating

• With GPS, accuracy is better than 73m in 75% of all measurements.

Network based locations was only better than 90m in 25% of all cases

Simon Wagner - User assisted analysis of cellular network structures 13

Wifi usage

• Smartphone users connect to wifi networks
• Decreases the load on the mobile networks
• Wifi connections are mostly encrypted
• Nearly no usage of WEP
• Mostly safe access via WPA
• Wifi-connected users don't deliver data about cellular

networks, but give insight into Internet connections of home networks

Wifi availability Wifi encryption

Simon Wagner - User assisted analysis of cellular network structures 14

Mobile data usage

• 3G technology is available most of the time
• 2G (GPRS, EDGE) still used
• 2G might be used to save energy

Simon Wagner - User assisted analysis of cellular network structures 15

Mobile data usage

• Speed was also measured
• However some incorrect data was reported and
• nly few speed measurements were available

Simon Wagner - User assisted analysis of cellular network structures 16

User tracking via global IP

• IP address travels with user
• GPRS core network handles the transition between base

stations

• Public IP address tells the network provider of the user: The

public IP belongs to an AS registered for the provider

User has IP address 2.213.203.149

User has still the same IP address here 100 km

Simon Wagner - User assisted analysis of cellular network structures 17

Future work – Possible Improvements

• Idea 1: Send phones through Germany to

collect informations along the route

• Our users stay mostly at the same area, so

sending a phone will deliver new informations

• Idea 2: Improve data collections: Start

measurements on users' phones only where no data has been collected

Simon Wagner - User assisted analysis of cellular network structures 18

Summary

• Reaching users all over the world is easy
• Programmers should implement both GPS and

network based location services

• Smartphone users use both Wifi and mobile

data connections to access the internet

• WEP is finally getting rare
• 3G is widely available, 2G is however still used

(possibly to save energy)

• Switching basestations still leaves you with the

same IP

Simon Wagner - User assisted analysis of cellular network structures 19

Thank you for your attention. Questions?