Transcript Part II : Connectivity Chapter 9: Opportunistic Networks

Ubiquitous
Computing
Max Mühlhäuser, Iryna Gurevych (Editors)
Part II : Connectivity
Chapter 9: Opportunistic Networks
Andreas Heinemann
Ubiquitous
Computing
Motivation
Short/medium range wireless communication technologies
capture the mass-market, e.g.
• Bluetooth enabled mobile phones
• WiFi enabled PDAs
• WiFi enabled mobile phones
 new network type called Opportunistic Networks
emerges
 based on spontaneous interaction and collaboration among
devices and users
Opportunistic Networks:
2
Ubiquitous
Computing
Application Example
At a computer science conference site, researchers from all around the world stay together for 2 –
3 days to discuss recent advances in their fields. Due to the limited time, each attendee tries to
make his stay as beneficial as possible, for example, by talking to colleagues during coffee breaks.
For novices in research there might be the question “Who should I talk to?” or “Which other
attendees are working on similar research problems?”
By carrying a Bluetooth enabled mobile phone, the device is able to communicate with nearby
devices carried by others in order to look for interesting conversational partners. Once the devices
have discovered a match in research interests, the devices notify their owners and the owners are
able to switch to a face-to-face communication due to the short communication range.
The devices might also exchange information, for example, paper reading lists, without user
notification. By this, each attendee would learn about what other researchers are currently working
on.
After the conference is over, this information is carried back home and the attendee might share
this information with colleagues at his research institute, again, by using his mobile phone and
without notice.
 Opportunistic Networks help to make people aware of each
other
 Support data dissemination similar to word-of-mouth
communication
Opportunistic Networks:
3
Ubiquitous
Computing
Underlying Ideas and Concepts
• User vicinity exploitation
• Profile based user interest expression
• Data dissemination
• Open and unrelated user group
• Unpredictable communication pattern
Opportunistic Networks:
4
Ubiquitous
Computing
A Definition for Opportunistic Networks
Definition (Opportunistic Network) An opportunistic network is a
network of wireless connected nodes. Nodes may be either
mobile or fixed. Communication range between two connected
nodes is within walking distance, i.e., 100–300 meters. The
network topology may change due to node mobility or node
activation and node deactivation. The nodes provide the
following functionality:
–
–
Node Discovery: A network node is able to discover other network
nodes in direct communication range.
One-hop Message Exchange: A node is able to send and receive
arbitrary data in form of a message to or from any other node in
direct communication range.
Definition (Opportunistic Network Node) An opportunistic network
node consists of a device with short-range wireless
communication capabilities. The device operates an
opportunistic network application that uses a data sharing
protocol for data dissemination. The data sharing protocol uses
i) node discovery and ii) one-hop message exchange.
Definition (Mobile Node) A mobile node (or node for short) consists
of a user carrying a mobile device that acts as an opportunistic
network node.
Definition (Information Sprinkler) An Information Sprinkler
(abbreviated IS) is a fixed opportunistic network node within
the network. It is a device placed at a dedicated location, thus
it is not mobile and not under direct user control. The
Information Sprinkler uses the same
data sharing protocol as other opportunistic network nodes.
Opp. Net.
Node
Mobile
Node
Infor.
Sprinkler
Opportunistic Networks:
5
Ubiquitous
Computing
Vertical Architecture
Opportunistic Networks:
6
Ubiquitous
Computing
MANETs for anonymous groups of humans?
• MANET = multi-hop ad-hoc network
• Sample application domains: Military, sensor networks, rescue scenarios
• Key characteristic: Common goal, strong relationship
• What is an incentive for B to route
messages between A and C?
• Why should A and C trust and rely on
node B for their communication?
?
C
A
B
Opportunistic Networks:
• One-hop communication to share information
– augmented with constrained propagation based on user profiles
– mimics word-of-mouth communication between humans
Opportunistic Networks:
7
Ubiquitous
Computing
P2P vs. MANET vs. Opp. Networks
Network
Type
Layer
Routing/Msg.
Forwarding
Focus
Node
Mobility
Network
Size
Community
Dynamics
Node
Relationship
P2P
Application
YES
NO
HIGH
HIGH
LOW
MANET
Network
YES
YES
LOW –
MEDIUM
MEDIUM
HIGH
Opp.
Network
Application
NO
YES
LOW
MEDIUM
LOW
Opportunistic Networks:
8
Ubiquitous
Computing
Opportunistic Networks Applications – Two Types
Active Collaboration
• exploits physical proximity of users in order to support a face-to-face
conversation
• device act as a link to the user
• Examples: Lovegety (Iwatani, 1998), SpotMe (Shockfish SA Switzerland,
2003), Nokia Sensor (Nokia, 2005)
Passive Collaboration
• disseminate data among nearby users without any user interaction
• digital form of word-of-mouth communiation
• Examples: Datta, Quarteroni, and Aberer (2004), Görgen et al. (2005),
Khelil, Becker, Tian, and Rothermel (2002)
Opportunistic Networks:
9
Ubiquitous
Computing
Opportunistic Network Example: iClouds
• Spontaneous one-hop network of humans
• Combines publish/subscribe with localized P2P networking
• Communication in user's vicinity
– no infrastructure needed
– spontaneous face-to-face meeting possible
• Digital items to share
– by interest
– using incentives
– no a-priori need for user's attention
• more info: http://iClouds.tk.informatik.tu-darmstadt.de
Opportunistic Networks:
10
Ubiquitous
Computing
Profile based data dissemination – Idea (iClouds)
Two basic data structures
• Information wish list
(iWish)
• Information have list
(iHave)
Opportunistic Networks:
11
Ubiquitous
Computing
Multi-Hop Information Dissemination (iClouds)
User A
iWish
User B
iHave
iWish
User C
iHave
iWish
iHave
user profile
most cases: to , L0
≠
t1 , L1
Opportunistic Networks:
12
Ubiquitous
Computing
Human Factors
Recall: Opportunistic Networks are formed by humans
carrying a personal device and potentially pass sensitive
information without notice.
Privacy Issues
 Q: How to protect a a user's privacy?
Incentive Issues
 Q: Why should a user contribute with a personal device to
a network? What is his benefit?
Opportunistic Networks:
13
Ubiquitous
Computing
Privacy – Degrees of User Identifiability
• Identity: A user that communicates with others and reveals any piece
of information that can be used to clearly identify him, is said to work
under his identity.
• Pseudonymity: This is the ability to prove a consistent identity without
revealing a user’s real identity, instead using a pseudonym.
(The harder it is to reveal the pseudonym of a user, the closer we are
to the state of not being identifiable at all, thus acting anonymously)
• Anonymity: Anonymity is the ability to remain unidentifiable within a
set. A user acts anonymously if it is impossible to reveal his identity.
Opportunistic Networks:
14
Ubiquitous
Computing
Privacy Preservation in iClouds
• Make use of dynamtic IDs during communication
• Idea
A
my ID is B
D
B
C
my ID is D
Typical network
stack
• Attention: All network layers need to be taken into account
Appl. layer
TCP/IP
802.11 WIFI
a number of self generated aliases
dynamic IP Addresses
dynamic MAC Addresses
Opportunistic Networks:
15
Ubiquitous
Computing
An Incentive Scheme Example
Basic Idea
• The incentive scheme rewards users (bearers) who partly
help to carry a piece of information from an information
producer to an information consumer.
Roles
• Information Producer
• Information Bearer
• Information Consumer
Opportunistic Networks:
16
Ubiquitous
Computing
Incentive Scheme Implementation: AdPASS
(Straub & Heinemann, 2004)
• AdPASS is a concrete Opportunistic Network
application based on iClouds
• Disseminates digital advertisements
according to user preferences (iWish/iHave)
• Bonus point reward for all people
carrying the ad to a buyer
Opportunistic Networks:
17
Ubiquitous
Computing
AdPASS: Participants & Communication Model
customer
A
B
C
bonus
2
5
3

vendors
disseminate
digital
customers
pass
on
the
ad
customer
returns
to
store
vendor
informs
mediator
customers
sync
their
bonus
ads
via
radio
to
customers
when
meeting
in
the
street
and
buys the product
about
points
pointsbonus
via internet




A

B
C
A
B
C
Opportunistic Networks:
18
Ubiquitous
Computing
Security Goals in AdPASS
Authentication
• assure that the information was issued by the claimed information
producer and not forged
Non-repudiation
• prevent an information producer from denying that he has issued a
certain piece of information
Integrity
• information integrity
• integrity of the bearer chain
Anonymity
• of information bearers in order to prevent an attacker from creating
user profiles
Opportunistic Networks:
19
Ubiquitous
Computing
Security Solutions in AdPASS (Overview)
Goal
Technique
Integrity
Digital signature operation
Authentication
Certificates
Non-Repudiation
Qualified signatures and
certificates
Anonymity
Multiple key pairs as aliases
Opportunistic Networks:
20
Ubiquitous
Computing
AdPASS: Integrity Protection of the Bearer Chain
• Make use of public key pairs (X+,X-)
–
–
+
X user alias
X for signature operation
P
10p
Information
Sender.: P+
Receiver.: A+
A
8p
B
2p
signed by PSender.: A+
Receiver.: B+
signed by A-
B's Attack: Remove A from chain
P
B
10p
10p
Information
Sender.: P+
Receiver.: B+
signed by P-
can't be forged by C without
knowledge of P
Opportunistic Networks:
21
Ubiquitous
Computing
•
•
•
•
•
•
•
Literature
Iwatani, Y. (1998). Love: Japanese Style. Retrieved February 2, 2007 from
http://www.wired.com/news/culture/0,1284,12899,00.html
Shockfish SA Switzerland. (2003). The SpotMe Homepage. Retrieved February 2, 2007
from http://www.spotme.ch Nokia. (2005). Nokia Sensor. Retrieved February 2, 2007
from http://www.nokia.com/sensor
Datta, A., Quarteroni, S., & Aberer, K. (2004). Autonomous Gossiping: A Self-Organizing
Epidemic Algorithm for Selective Information Dissemination in Wireless Mobile Ad-Hoc
Networks. Lecture Notes in Computer Science, 3226, 126–143.
Görgen, D., Frey, H., & Hutter, C. (2005). Information Dissemination Based on the EnPassent Communication Pattern. In Kommunikation in verteilten systemen (kivs 2005)
(pp. 129–141).
Khelil, A., Becker, C., Tian, J., & Rothermel, K. (2002). An Epidemic Model for
Information Diffusion in MANETs. In Mswim ’02: Proceedings of the 5th acm international
workshop on modeling, analysis, and simulation of wireless and mobile systems (pp. 54–
60). New York, NY, USA: ACM Press.
Straub, T., & Heinemann, A. (2004). An Anonymous Bonus Point System For Mobile
Commerce Based On Word-Of-Mouth Recommendation. In L. M. Liebrock (Ed.), Applied
computing 2004. proceedings of the 2004 acm symposium on applied computing (pp. 766–
773). New York, NY, USA: ACM Press.
Heinemann. A (2007) Collaboration in Opportunistic Networks Ph.D. Thesis, University of
Technology, Darmstadt, 2007. http://elib.tu-darmstadt.de/diss/000834
Opportunistic Networks:
22