Thesis Presentation

Supervisor: Prof. Jörg Ott

Student: Junxi Yin

Espoo, 11.21.2012

Topic

Mobile Implementation of Floating Content Service

Outline

• Introduction• Delay-tolerant Networking• Floating Content Service• Application Design and Implementation• Test and Evaluation• Conclusion and Future Work• References• Appendix

Background

Scenarios where traditional Internet is not working well

Flea market

Travelling

Rural areas

Supermarket

Developing countries

…

Background

Why DTN in mobile?

• Development of DTN technology• Growing mobile market (Cheaper devices)• Growing usage of mobile devices• Growing usage of location service• Open source sw and tools

Objective

• 1. Theoretical studies

• 2. Simulations

• 3. Implementation and test

DTNDTN=Delay-Tolerant Networking [1]

• Designed for extreme conditions • Protocol Data Unit - Bundle • Convergence Layer • Custody Transfer• Store-and-Forward

D

S

Store-and-Forward

Source Node

Custodian A Destination Node

Custodian B

Persistent Storage

Forward Forward Forward

Persistent Storage

Data is stored throughout the network and forwarded when a better chance comes in hopes that it will eventually reach its destination – Opportunistic Communication

Floating Content Service

A DTN-based localized information sharing service exclusively dependent on mobile phones and other mobile devices. [2]

As long as there are enough supportive mobile devices around in the anchor zone to replicate and store a piece of content, it floats.

ar

Node with a message

Node without a message

Replication Probability

r a

1

0

p

d

Replication Probability According to Distance from Message Center

Since replication is based purely on the location of nodes, in a simple case, every node in the anchor zone should have a copy of the item.

In practice, the replication works as follows.

SimulationsONE Simulator=Opportunistic Network Environment simulator

-Written in Java-Agent-based, discrete event, network simulator-Support Mobility Models-Several built-in DTN routing algorithms [3]

More Information:

http://www.netlab.tkk.fi/tutkimus/dtn/theone/

Location Error ConsiderationRadio range: 50, Nodes: 126, ttl=3600, r=500, a=500, size=minimal, buffer=5M

Mean of all anchor pointsMean of anchor points with n>100 messages

Weighted average

Fraction of messages available <= x

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Lifetime as fraction of TTL

1

0

Radio range: 50, Nodes: 126, ttl=3600, r=500, a=500, size=minimal, buffer=5M

Mean of all anchor pointsMean of anchor points with n>100 messages

Weighted average

Fraction of messages available <= x

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Lifetime as fraction of TTL

1

0

Performance of Floating Content Service without Location Error Performance of Floating Content Service with Location Error

Please refer to ”Floating Content: Information Sharing in Urban Areas” to get more details about the performance of Floating Content service

A Closer LookRadio range: 50, Nodes: 126, ttl=3600, r=500, a=500, size=minimal, buffer=5M

Mean of all anchor pointsMean of anchor points with n>100 messages

Weighted average

Fraction of messages available <= x

0

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Lifetime as fraction of TTL

-0.1

0.1

Compared floating performance as a function of lifetime

Conclusion:

The location error has very limited effect to the performance of Floating Content application

System Design

User Interface

Service Engine

create

save

created data repository

received data repository

saved data repository

Output

Input

get get

get

subscribe

System Implementation

Open source tools, no costs at all.

• Java• Android• Floating Content API• SQLite Database• XML

Action Flow

Subscribed?

Message received

No

Discard

Yes

Data stores in received data

repository

Message sent out

TTL expired?

Message created

Data stores in created data

repository

No

Yes

Delete

TTL expired?No

Yes

Delete

Data stores in saved data repository

Application starts

Service engine starts

Save

Screenshots

Test Results

Transfer Time (Second)

0 1 2 3 4 5 6 7 8 9 10

Size of Attachement File (MB)

4

0

1

2

3

9

5

6

7

8

12

10

11

Transfer Speed of Floating Content Service between Two Devices

The transmission speed using Floating Content service is similar with the actual Wi-Fi speed between only two devices.

Future Work

• Deploy a systematic testbed to better evaluate the performance of Floating Content service in a real world environment

• Better user interface design

References

[1] Forrest Warthman, “Delay-Tolerant Networks (DTNs), A Tutorial”, May 2003.

[2] J. Kangasharju, J. Ott, and O. Karkilahti, “Floating Content: Information Availability in Urban Environments,” in Proc. of IEEE Percom 2010, Work in Progress session, March 2010.

[3] A. Keränen, J. Ott, and T. Kärkkäinen, “The ONE Simulator for DTN Protocol Evaluation”, in SIMUTools ’09:Proceedings of the 2nd International Conference on Simulation Tools and Techniques, NewYork, NY, USA: ICST, 2009.

Thank You !

• Junxi Yin• junxi.yin@aalto.fi