Transcript Specialized Ad Hoc Networks for Emergency Response Applications

CS 6910 – Pervasive Computing
Section 0.B:
Opportunistic Networks:
Specialized Ad Hoc Networks
for Emergency Response Applications
Dr. Leszek Lilien
WiSe Lab (Wireless Sensornet Laboratory)
http://www.cs.wmich.edu/wsn
Department of Computer Science
Western Michigan University
Slides are © 2007 by Leszek T. Lilien
Requests to use original slides for non-profit purposes will be gladly granted upon a written request.
Outline



Part 1. Motivation for Specialized Ad Hoc Networks
Part 2. Analogy to a Human Emergency Response
Team
Part 3. Opportunistic Networks: A New Type of
Specialized Ad Hoc Networks

Part 4. Related Research and Research Challenges

Part 5. Conclusions

Part 6. Current and Future Work
2
Part 1.
Motivation for Specialized Ad Hoc Networks


Homeland Security
 One of the crucial challenges facing the USA
today
 Among its 6 mission areas is
Emergency Preparedness and Response (EPR)
EPR deals with:
 Natural disasters
 Man-made disasters
(incl. accidents, terrorist attacks)
[Natl. Strategy for Homeland Security, July 2002]

Mobile ad hoc networks (MANETs) proposed for EPR
[Haas, 1999]
3
Motivation for Specialized Ad Hoc Networks (2)


MANETS are not quite a natural match for EPR
 E.g., human rescue teams can find and mobilize
as their helpers local firemen, police, National
Guard or even regular citizens
 No analogous capability of MANETs to find and
“mobilize” devices/networks
Let’s find or define a specialization (a subclass) of ad
hoc networks more suitable for EPR applications
 A new paradigm and a new technology to
improve effectiveness & efficiency of EPR
4
This page left blank intentionally.
5
Part 2. Analogy to
a Human Emergency Response Team

Important characteristics of a human rescue team

Starts as a seed team (a small group of people)

The seed team grows opportunistically during its
operations
a) Some people can always be ordered to join
 Police, firemen, National Guard or Army Reserve
personnel
b) Anybody can be ordered to join only in life-ordeath situations
 Legally required to help saving lives or critical
resources
c) Anybody can be asked to join in other
situations
6
Analogy to a Human Emergency Response Team (2)


Who is ordered or invited to join?

In some situations any extra pair of hands can
help

In others only highly qualified people (e.g., doctors
and nurses) are ordered/asked to join
Human helper types

Highly prepared


Somewhat prepared


E.g., National Guard, Army Reserve or state militias

Signed up for service

Undergo training in preparation for duty (even unforeseen)
E.g., people that volunteer for first aid courses

Become pretty valuable helpers in emergencies
Not prepared at all
7
Analogy to a Human Emergency Response Team (3)

Benefits of the opportunistic growth of the rescue
team
 Opportunistic leveraging of all kinds of skills and
resources that new helpers can bring

Obtaining a lot of help effectively and efficiently – even
for free
8
Analogy to a Human Emergency Response Team (4)
Analogous critical requirements for ad hoc network
specialized for EPR in the priority order:
1) Minimal starting configuration – a pre-configured seed for
EPR operations

Analogy to the seed rescue team
2) High interoperability in terms of communication of diverse
devices or networks (Wired/WiFi, Bluetooth, satellite, ham radios, WiMAX, ...)
Analogy to a rescue team’s ability to contact different people, individually
or via organizations
3) Highly heterogeneous software (& hardware)
Analogy to heterogeneity of rescue teams in terms of members’ skills,
communication and other equipment, and other resources
4) Harvesting of diverse resources as needed
Analogy to finding people with different skills, equipment, and other
resources
5) Persistent connectivity once it is established
Analogy to being able to contact (maybe via a chain of others) members of the
expanded team, including all helpers
9
Analogy to a Human Emergency Response Team (5)


If there is no specialized ad hoc network (AHN) matching the
requirements, then:
 Define a new specialized AHN paradigm
 Invent a new specialized AHN technology
No known specialized AHN matches the requirements
Considered AHNs:
 Mobile ad hoc networks (MANETs)
 Mesh networks
 P2P systems
 Sensor networks
 Spontaneous networks (in the narrow sense, cf. [Feeney et al. 2001])
=> Need:
- a new specialized AHN paradigm
- a new specialized AHN technology
Note: Other (than oppnets) specialized AHNs for emergencies
are Incident Area Networks (IANs)
10
This page left blank intentionally.
11
Part 3. Opportunistic Networks:
A New Type of Specialized Ad Hoc Networks

Paradigm of opportunistic networks (oppnets)


Based on the analogy to operations of human rescue
teams
Goals of oppnets:
 Opportunistic growth
 Opportunistic leveraging of resources aiming at
(among others):
 Bridging diverse communication media
 Offloading computations to additional
platforms
 Integrating independent sensing systems
(enhancing their sensing capabilities)
12
Startup: Seed Oppnet

Oppnet deployed as a seed oppnet
Link to
the World
Seed Nodes
Controller
(distributed)

Seed oppnet localizes its nodes & self-configures
13
Growth: Expanded Oppnet

Seed oppnet grows into an expanded oppnet by:


Finding candidate helpers
Selecting candidates - they are ordered/asked to join

Add communication, computing, sensing, storage, other resources
Satellite
Link to
the World
Appliance
Cellphone
Tower
(refrigerator)
Seed Nodes
Controller
(distributed)
WiMAX
Overturned
Vehicle
(with OnStar, VAN)
Computer Network 1414
Summary of Oppnet Activities

Discovering & identifying candidate helpers

Contacting selected candidates

Inviting or ordering candidates to join

Admitting & integrating helpers that join oppnet

Offloading tasks to helpers

Determining useful colaborative functionalities

Managing offloaded tasks

Cleaning up & releasing each helper when no longer
needed

> Need research for each of theses activities <

Think which is most interesting to you as a topic for your
CS6910 Team Project
15
Oppnet Reserve


A challenge in oppnet growth
 Must discover candidate helpers, then contact selected
ones
 Difficult without facilities provided by candidates or
infrastructure
A solution: oppnet reserve — facilitating discovery/contacting
 Analogy to Army / Navy / Air Force Reserve, etc.
 Volunteer helpers sign up for oppnet reserve
 Maybe for some incentives (moral, financial, etc.)
 Volunteers „trained” for active duty
 Install facilities that make them easier to detect and
contact by oppnets


E.g., install (future) standard oppnet protocols
 Matched to their capabilities: heavy-, medium- and
lightweight
Available for active oppnet duty whenever must/can help
16
Oppnet Reserve (2)

Oppnet reserve is analogous to having highly
prepared human helpers
(Recall the classification of human helpers into: highly prepared,
somewhat prepared, unprepared)

Oppnet reserve is not required but very helpful

Having highly prepared human helpers is not required
either
17
What Oppnet Is … / What Oppnet Is Not …
(this slide added after lecture)

Oppnet is …

… a specialized (application-specific or application-classspecific) ad hoc network

… an opportunistic growth network

… an opportunistic use of resources network /
an opportunistic use of services network



Spectrum, connectivity, services, …
… a privacy-preserving & secure technology
Oppnet is not …

… a “generic” ad hoc network

… a mesh network

… a grid computing system

… a peer-to-peer (P2P) system

… a hybrid network

…
18
Basic Oppnet Categories


2 major oppnet categories:

Benevolent oppnets

Malevolent oppnets
Corresponding oppnets scenarios:

Benevolent oppnet scenario:
„Citizens Called to Arms”

Malevolent oppnet scenario:
„Bad Guys Gang Up”
19
Benevolent Oppnet
Scenario: „Citizens
Called to Arms” (1)

Seed oppnet deployed
after an earthquake (unpredictable emergency)

Seed is ad hoc wireless network with very powerful
nodes


20
More energy, computing and communication resources
Seed tries to detect candidate helpers

For help in damage assessment and disaster recovery

Uses any available detection method — including:

Cellphone- or radio-based detection



Including Software Defined Radio (SDR) & Cognitive Radio
Searching for nodes using the IP address range for the
affected geographic area
AI-based visual detection (next slide)
20
Benevolent Oppnet
Scenario: „Citizens
Called to Arms” (2)

Example:

Helper 1
(supervisor of
security cameras) monito-






21
ring a surveillance net views an overturned car
Helper 2 (pattern recognition specialist) detects an overturned
car
Helper 3 (image analysis specialist) asked to recognize car’s
license plate
Helper 4 (DB manager) finds that the car has OnStar link
Helper 5 (OnStar agent) contacts VAN (Vehicle Area Network)
and BANs (Body Area Networks) on or within bodies of car
occupants via OnStar infrastructure
Helper 6 (vital sign evaluator) evaluates obtained info
Helper 7 (rescue dispatcher) decides if/when rescuers
should be dispatched
21
Benevolent Oppnet
Scenario: „Citizens
Called to Arms” (3)
22


Oppnet selects optimal subset of detected nodes

Inviting devices, clusters & entire networks

Helpers for communicating, sensing, computing
Using „hidden” capabilities, e.g. for sensing:

Desktop can „sense” presence of a potential victim at
its keyboard

Cellphones can „sense” location

Even ones w/o GPS can be triangulated
22
Benevolent Oppnet
Scenario: „Citizens
Called to Arms” (4)

Using „hidden”
emergency functionalities

Oppnet contacts 2 independent sensornets (SNs):
water infrastructure control SN /
public space surveillance SN

SNs ordered to abandon normal functions & help in
rescue & recovery operations

Water infrastructure SN (with multisensor capabilities,
under road surfaces) — ordered to sense vehicular
movement and traffic jams

Public space surveillance SN — ordered to search
for images of human victims
23
23
Malevolent Oppnet
Scenario: „Bad
Guys Gang Up” (1)

Scenario 1 — Terrorists
create apparently
24
harmless weather monitoring sensornet (SN):

SN becomes a seed of a malevolent opportunistic SN

SN exploits other nodes from many other networks (w/o
revealing its true goals)

“Critical mass” of the opportunistic SN is reached
geographical spread and sensing capabilities)

(in terms of
SN waits for wind patterns that can speed up spread of
poisonous chemicals

Collected data used to decide when to start chemical attack
24
Malevolent Oppnet Scenario:
„Bad Guys Gang Up” (2)

Scenario 2 — network at home starts spying on you:

Becomes a seed oppnet

Exploits other devices/nets to collect all info on you:










From your fridge (& RFID-equipped food packaging): what/when you
eat
From your computer: keylogs your passwords, sensitive data
From your cellphone: who you call & when
From your networked camera: what photos you take
From your home security surveillance system: your private
images
Cyberfly with camera eyes and microphone ears
...
Huge privacy problem! / Huge security problem!
Controls to counteract malevolent oppnets badly
needed
25
NOTE:
The following slides not required for exam.
(Provided as a potential help for project
selection and project work.)
26
Part 4. Related Research
& Research Challenges

Examples of Related Research
 Interoperability

Among wireless networks - active research area


Among WANs, MANs, LANs, PANs (Personal Area Networks), etc.
Among wired & wireless nets - much less research

Ambient Networks (big European Union project, next-generation Internet—for
2015/2020, smaller networks able to compose themselves into bigger ones)

Localization & self-organization


Network growth



P2P systems – search for peers in unstructured systems
Trojan Horses - agents spreading in search for helpers
Integrating and managing heterogeneous systems, incl.
data integration & aggregation


MANETs / Sensornets
Grid systems / MANETs / Sensornets
Other

…
27
Research Challenges in Basic Operations

Challenges in seed oppnet deployment


E.g., localization, self-configuration, adaptability
Challenges in detecting helper systems

E.g., define primitives to detect candidates, identify and
categorize them, evaluate and classify them (e.g., based on
dependability and usefulness)

Challenges in inviting & admitting candidate helpers


E.g., select candidates to invite, develop protocols for
candidates to accept or reject invitation, devise primitives
/methods to manage expanded oppnets
Etc., etc. for remaining oppnet primitives
More:
Leszek Lilien, Z. Huma Kamal, and Ajay Gupta, "Opportunistic Networks:
Research Challenges in Specializing the P2P Paradigm,” Proc. 3rd
International Workshop on P2P Data Management, Security and Trust
(PDMST’06), Kraków, Poland, September 2006.
28
Research Challenges in Privacy & Security (1)
1) Privacy challenges in oppnets

Privacy is critical

Oppnets are pervasive systems
 Must face all critical privacy challenges inherent to
pervasive computing

Privacy is a „make it or break it” issue for pervasive
computing =>
Privacy is a „make it or break it” issue for oppnets

Basic privacy protection goals in oppnets

Protect helper resources from the host oppnet

Protect oppnet from its helpers

Protect environment from privacy violations by
oppnet

Also from malevolent oppnets
29
Research Challenges in Privacy & Security (2)
2) Security challenges in oppnets

Many have privacy aspects/components
More:
L. Lilien, Z. H. Kamal, V. Bhuse, and A. Gupta, "Opportunistic
Networks: The Concept and Research Challenges in Privacy and
Security," book chapter in: "Mobile and Wireless Network Security and
Privacy," ed. by K. Makki et al., Springer Science+Business Media,
Norwell, Massachusetts, 2007 (to appear)
– pre-publication version available upon request
30
This page left blank intentionally.
31
Part 5. Conclusions

High-payoff potential for the oppnet initiative
1) Social and economic benefits
 Including reduction of human suffering & loss of life
2) Technological benefits
3) Research benefits
4) Educational benefits
-- Details below -32
Part 5. Conclusions – cont.2
1) Social & Economic Impacts

Impacts on Emergency Preparedness and
Response operations in Homeland Security (HS) —
current app focus

Tremendous leveraging potential in emergencies





A wealth of freely available resources
Reduction of human suffering & loss of life
Increasing safety & efficiency of the first responders
Impacts on other applications both in HS and
outside HS
[SKIP:] Economic impacts


Technology transfer & commercialization
Benefits for the computer industry


Production of software / networking equipment
Benefits for other industries


Enhancing many products with standard oppnet interfaces
E.g., the auto industry: cars and trucks as oppnet platforms
33
Part 5. Conclusions – cont.3
2) Impacts on Technology



Advancing the network and pervasive computing know-how

Development of the innovative oppnet technology

Enhancing network/pervasive applications by use of
oppnet technologies

Enabling new network/pervasive application niches we
can not even foresee
Advancing other areas of technology (not only computing)

A side effect of oppnet developments
[SKIP:] Technology impacts speeded up & enhanced by the
planned technology transfer plus commercialization
activities
34
Part 5. Conclusions – cont.4
[SKIP:] 3) Impacts on Research
[SKIP:] 4) Impacts on Education
35
This page left blank intentionally.
36
Part 6. Current and Future Work


>> Search for good uses for oppnets in all kinds of
application areas <<
Search for applications that ...

... can start with a seed

... need/have high interoperability

... need/have highly heterogeneous software

... can benefit from leveraging diverse
resources of helpers

... can maintain persistent connectivity once it
is established

...
37
Part 6. Current and Future Work – cont.

Building an oppnet prototype
 Goal: Proof of concept


To demonstrate technical prowess & economic
benefits
Designing oppnet architecture

With its associated components:


Methods, protocols, and algorithms
Oppnet prototype implementation

For stimulation and feedback

Necessary for fine-tuning oppnet design
38
Publications, Etc.

Selected Publications
(cf. http://www.cs.wmich.edu/wsn/project_oppnet.html)
L. Lilien, A. Gupta, and Z. Yang, "Opportunistic Networks for Emergency Applications and
Their Standard Implementation Framework," Proc. The First International Workshop on
Next Generation Networks for First Responders and Critical Infrastructure (NetCri07),
New Orleans, Louisiana, April 11-13, 2007 (to appear)
L. Lilien, Z. H. Kamal, V. Bhuse, and A. Gupta, "Opportunistic Networks: The Concept and
Research Challenges in Privacy and Security," book chapter in: "Mobile and Wireless
Network Security and Privacy," ed. by K. Makki et al., Springer Science+Business Media,
Norwell, Massachusetts, 2007 (extended version of the WSPWN 2006 paper; to appear)
L. Lilien, Z. H. Kamal and A. Gupta, "Opportunistic Networks: Research Challenges in
Specializing the P2P Paradigm," Proc. 3rd International Workshop on P2P Data
Management, Security and Trust (PDMST'06), Kraków, Poland, September 4-8, 2006, pp.
722-726.
L. Lilien, Z. H. Kamal, V. Bhuse and A. Gupta, "Opportunistic Networks: The Concept and
Research Challenges in Privacy and Security," Proc. International Workshop on Research
Challenges in Security and Privacy for Mobile and Wireless Networks (WSPWN 2006),
Miami, Florida, March 2006, pp. 134-147.
B. Bhargava, L. Lilien, A. Rosenthal and M. Winslett, "Pervasive Trust," IEEE Intelligent
Systems, vol. 19(5), Sep./Oct.2004, pp. 74-77 (just the first brief mention of the oppnet
idea, in the form of malevolent opportunistic sensor networks).
39
Publications, Etc. – cont. 1

Selected Presentations
L. Lilien, "Opportunistic Networks: Research Challenges in Specializing the P2P
Paradigm," 3rd International Workshop on P2P Data Management, Security and
Trust (PDMST'06), Kraków, Poland, September 4-8, 2006
L. Lilien, "Opportunistic Networks: Specialized Ad Hoc Networks for Emergency
Response Applications," presented for Distributed Systems Research Group,
Department of Computer Science, AGH University of Science and Technology,
Krakow, Poland, May 22, 2006
L. Lilien, "Opportunistic Networks: Specialized Ad Hoc Networks for Emergency
Response Applications," presented for Section of Information Technology,
Institute of Telecomputing, Cracow University of Technology, Kraków, Poland,
May 23, 2006
L. Lilien, "Developing Specialized Ad Hoc Networks: The Case of Opportunistic
Networks," Workshop on Distributed Systems and Networks (in conjunction
with WWIC’06), Bern, Switzerland, May 9, 2006
L. Lilien, "Opportunistic Networks: The Concept and Research Challenges in
Privacy and Security," International Workshop on Research Challenges in
Security and Privacy for Mobile and Wireless Networks (WSPWN 2006), Miami,
Florida, March 15-16, 2006
L. Lilien, Z.H. Kamal and A. Gupta (in cooperation with V. Bhuse and Z Yang),
"Opportunistic Networks: The Concept and Research Challenges," Department
of Computer Science, Western Michigan University, Kalamazoo, Michigan,
February 9, 2006
40
Publications, Etc. – cont. 2

Selected Posters
Leszek Lilien, Zille Huma Kamal, A. Gupta, V. Bhuse and Z. Yang, "Opportunistic
Networks," 3rd International Conference on Networked Sensing Systems, Chicago, IL,
May 31 - June 2 2006.
Leszek Lilien, Zille Huma Kamal, Vijay Bhuse and Ajay Gupta, "Opportunistic Networks
and Their Privacy and Security Challenges," The Seventh Annual CERIAS Information
Security Symposium - "Negotiating Trust: Security, Privacy, Risk," CERIAS, Purdue
University, West Lafayette, IN, March 21, 2006.
Leszek Lilien and Ajay Gupta, "Opportunistic Networks for Emergency Preparedness and
Response," 4th Annual e-Enterprise Conference, PHSI/RCHE, Purdue University, West
Lafayette, IN, March 20, 2006.
41
WiSe Lab Experience in Ad Hoc Networks –
Selected Projects Since 1/03
















Designing of WiSe Security Protocols: DSPS
Location Tracker Using Motes (*)
RHS: Remote Home Surveillance (*)
Directed Diffusion: Attacks & Countermeasures
Improving the Accuracy of Mote Measurements
by Using Neural Networks
SOMS: Smart Occupancy Monitoring System Using Motes (*)
Comparative Study of Network Simulators
Collaborative Image Processing (*)
DENSe: a Development Environment for Networked Sensors
Incorporating Mobile-ware in Distributed Computations / Grids (*)
Extending the ns-2 Simulator to Satellite and WCN Simulations
Smart Antennas for WCNs
Energy Efficient MAC Protocols for IEEE 802.11x
A Wireless Security Testing System (*)
Mobile and Self-Calibrating Irrigation System
Collective Communications for Sensornets (*)
* Results directly useful for oppnets
42
This page left blank intentionally.
43