How Can Botnets Cause Storms? Understanding the Evolution ...
Transcript of How Can Botnets Cause Storms? Understanding the Evolution ...
![Page 1: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/1.jpg)
How Can Botnets Cause Storms? Understandingthe Evolution and Impact of Mobile Botnets
Zhuo Lu†, Wenye Wang†, and Cliff Wang‡
† Department of Electrical and Computer EngineeringNorth Carolina State University, Raleigh NC, US.
‡ Army Research OfficeResearch Triangle Park NC, US.
Apr., 2014
1 / 25
![Page 2: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/2.jpg)
Outline
1 MotivationMobile Applications, Malware and BotnetsResearch Issues and Our Focus
2 PreliminariesNetwork and Attack ModelsProblem Formulation
3 ResultsBotnet PropagationMobile Botnet Impact
4 Conclusion
2 / 25
![Page 3: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/3.jpg)
Outline
1 MotivationMobile Applications, Malware and BotnetsResearch Issues and Our Focus
2 Preliminaries
3 Results
4 Conclusion
3 / 25
![Page 4: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/4.jpg)
Smart Phone, Mobile Malware, and Mobile Botnet
Smart Phones
Powerful hardware, mobile operating systems, mobile APPs.
Mobile malware has come into practice.
4 / 25
![Page 5: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/5.jpg)
Smart Phone, Mobile Malware, and Mobile Botnet
Smart Phones
Powerful hardware, mobile operating systems, mobile APPs.
Mobile malware has come into practice.
The Threat of Mobile Botnets
Mobile botnet: A collection of malware infected nodes able toperform coordinated attacks.
Ikee.B in 2009
Android.Bmaster in 2011.
4 / 25
![Page 6: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/6.jpg)
Smart Phone, Mobile Malware, and Mobile Botnet
Smart Phones
Powerful hardware, mobile operating systems, mobile APPs.
Mobile malware has come into practice.
The Threat of Mobile Botnets
Mobile botnet: A collection of malware infected nodes able toperform coordinated attacks.
Ikee.B in 2009
Android.Bmaster in 2011.
[Traynor ’09]: a botnet with sufficiently many infected phonesis able to disrupt regional cellular services.
4 / 25
![Page 7: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/7.jpg)
How Mobile Botnets Propagate in the Network
The ways that a botnet propagates in mobile networks
Centralized propagation: SMS/MMS, APPs in the market.Becoming harder and harder.
Mobile-to-mobile/Proximity infection: More stealthy!
mobility trac
e
infected node
susceptiblesusceptible
infected infected
…
…
…
Existing malware adopting proximity infection.E.g., Mabir, Lansco and CPMC.
5 / 25
![Page 8: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/8.jpg)
How Mobile Botnets Propagate in the Network
The ways that a botnet propagates in mobile networks
Centralized propagation: SMS/MMS, APPs in the market.Becoming harder and harder.
Mobile-to-mobile/Proximity infection: More stealthy!
mobility trac
e
infected node
susceptiblesusceptible
infected infected
…
…
…
Existing malware adopting proximity infection.E.g., Mabir, Lansco and CPMC.
5 / 25
![Page 9: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/9.jpg)
How Mobile Botnets Propagate in the Network
The ways that a botnet propagates in mobile networks
Centralized propagation: SMS/MMS, APPs in the market.Becoming harder and harder.
Mobile-to-mobile/Proximity infection: More stealthy!
mobility trac
e
infected node
susceptiblesusceptible
infected infected
…
…
…
Existing malware adopting proximity infection.E.g., Mabir, Lansco and CPMC.
5 / 25
![Page 10: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/10.jpg)
How Mobile Botnets Propagate in the Network
The ways that a botnet propagates in mobile networks
Centralized propagation: SMS/MMS, APPs in the market.Becoming harder and harder.
Mobile-to-mobile/Proximity infection: More stealthy!
mobility trac
e
infected node
susceptiblesusceptible
infected infected
…
…
…
Existing malware adopting proximity infection.E.g., Mabir, Lansco and CPMC.
5 / 25
![Page 11: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/11.jpg)
Research Question and Issues in the Literature
Question?
Can Mobile Malware via Proximity Infection Cause Storms?
6 / 25
![Page 12: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/12.jpg)
Research Question and Issues in the Literature
Question?
Can Mobile Malware via Proximity Infection Cause Storms?
Answers
Yes ([Carettoni’07, Yan’09, Wang’09]): Epidemic modelingand experiments
Infection storm: More and more nodes get infected as timegoes.
No ([Husted’11]): Simulations in realistic mobile scenarios.
Limited infection: the number of infected devices is limitedwith the relatively low vulnerability ratio.
6 / 25
![Page 13: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/13.jpg)
Research Question and Issues in the Literature
Question?
Can Mobile Malware via Proximity Infection Cause Storms?
Answers
Yes ([Carettoni’07, Yan’09, Wang’09]): Epidemic modelingand experiments
Infection storm: More and more nodes get infected as timegoes.
No ([Husted’11]): Simulations in realistic mobile scenarios.
Limited infection: the number of infected devices is limitedwith the relatively low vulnerability ratio.
Somewhat discrepant results in the literature.
Why: Node density, mobility, vulnerability ratio?
6 / 25
![Page 14: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/14.jpg)
Research Question and Objective
Research Question
How to model the botnet propagation and impact in mobilenetworks?
Objectives
1 Characterize how fast a mobile botnet propagates.
2 Investigate the denial-of-service impact of such a botnet.
7 / 25
![Page 15: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/15.jpg)
Research Question and Objective
Research Question
How to model the botnet propagation and impact in mobilenetworks?
Objectives
1 Characterize how fast a mobile botnet propagates.
2 Investigate the denial-of-service impact of such a botnet.
7 / 25
![Page 16: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/16.jpg)
Outline
1 Motivation
2 PreliminariesNetwork and Attack ModelsProblem Formulation
3 Results
4 Conclusion
8 / 25
![Page 17: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/17.jpg)
Network Model
A hybrid network: infrastructure and mobile nodes.
mobile nodes
moving around
malware-infected nodes
also moving around
Infrastructure
node
transmission
radius r
...
... ...
......
...
...
...
...
...
...
(a)
(b)
transmission range r, mobile node density λ, network bandwidth B
9 / 25
![Page 18: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/18.jpg)
Propagation Model: Proximity Infection
How to propagate malware from one node to another?
mobility trace
infected node
infected
susceptible
1 One is infected, another is vulnerable (vulnerability ratio κ).
2 Two nodes are in each other’s transmission range (r).
3 Meeting time > threshold.
10 / 25
![Page 19: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/19.jpg)
Propagation Model: Proximity Infection
How to propagate malware from one node to another?
mobility trace
infected node
infected
susceptible
1 One is infected, another is vulnerable (vulnerability ratio κ).
2 Two nodes are in each other’s transmission range (r).
3 Meeting time > threshold.
10 / 25
![Page 20: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/20.jpg)
Propagation Model: Proximity Infection
How to propagate malware from one node to another?
mobility trace
infected node
infected
susceptible
1 One is infected, another is vulnerable (vulnerability ratio κ).
2 Two nodes are in each other’s transmission range (r).
3 Meeting time > threshold.
10 / 25
![Page 21: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/21.jpg)
Propagation Model: Proximity Infection
How to propagate malware from one node to another?
mobility trace
infected node
infected
susceptible
1 One is infected, another is vulnerable (vulnerability ratio κ).
2 Two nodes are in each other’s transmission range (r).
3 Meeting time > threshold.
10 / 25
![Page 22: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/22.jpg)
Mobility Model: Generic Mobility
Realistic mobility always incurs spatial heterogeneity.
Network
11 / 25
![Page 23: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/23.jpg)
Mobility Model: Generic Mobility
Realistic mobility always incurs spatial heterogeneity.
Network
home pointmobility radius α
distribution Ψ
11 / 25
![Page 24: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/24.jpg)
Problem Formulation and Performance Metric
Botnet S(t): the set of all infected nodes at t.
Question: What is the botnet size |S(t)| at time t?
12 / 25
![Page 25: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/25.jpg)
Problem Formulation and Performance Metric
Botnet S(t): the set of all infected nodes at t.
Question: What is the botnet size |S(t)| at time t?
12 / 25
![Page 26: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/26.jpg)
Problem Formulation and Performance Metric
Botnet S(t): the set of all infected nodes at t.
Question: What is the botnet size |S(t)| at time t?
12 / 25
![Page 27: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/27.jpg)
Outline
1 Motivation
2 Preliminaries
3 ResultsBotnet PropagationMobile Botnet Impact
4 Conclusion
13 / 25
![Page 28: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/28.jpg)
Theoretical Results
Theorem: Mobile Botnet Propagation
If the value of κλ(2α + r) is sufficiently large, we have a botnetpropagation storm: the average botnet size E|S(t)| = Θ(t2).
|S(t)|
quadratic growth
Otherwise, we have limited propagation: E|S(t)| = Θ(1).
14 / 25
![Page 29: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/29.jpg)
Theoretical Results
Theorem: Mobile Botnet Propagation
If the value of κλ(2α + r) is sufficiently large, we have a botnetpropagation storm: the average botnet size E|S(t)| = Θ(t2).
|S(t)|
quadratic growth
Otherwise, we have limited propagation: E|S(t)| = Θ(1).
|S(t)|
limited growth
14 / 25
![Page 30: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/30.jpg)
Theoretical Results
Theorem: Mobile Botnet Propagation
If the value of κλ(2α + r) is sufficiently large, we have a botnetpropagation storm: the average botnet size E|S(t)| = Θ(t2).Otherwise, we have limited propagation: E|S(t)| = Θ(1).
Direct Indications
1 Fastest rate of proximity infection: quadratic growth.
Internet botnets: exponential growth.
14 / 25
![Page 31: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/31.jpg)
Theoretical Results
Theorem: Mobile Botnet Propagation
If the value of κλ(2α + r) is sufficiently large, we have a botnetpropagation storm: the average botnet size E|S(t)| = Θ(t2).Otherwise, we have limited propagation: E|S(t)| = Θ(1).
Direct Indications
1 Fastest rate of proximity infection: quadratic growth.
Internet botnets: exponential growth.
2 κλ(2α + r) is the key
density λ, mobility radius α, transmission range r.
14 / 25
![Page 32: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/32.jpg)
Theoretical Results
Theorem: Mobile Botnet Propagation
If the value of κλ(2α + r) is sufficiently large, we have a botnetpropagation storm: the average botnet size E|S(t)| = Θ(t2).Otherwise, we have limited propagation: E|S(t)| = Θ(1).
Direct Indications
1 Fastest rate of proximity infection: quadratic growth.
Internet botnets: exponential growth.
2 κλ(2α + r) is the key
density λ, mobility radius α, transmission range r.
Practical scenario: density λ and transmission range r fixedSufficient mobility always triggers the Θ(t2) infection!
14 / 25
![Page 33: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/33.jpg)
Experimental Evaluation: Setups
Mobility traces: EPFL/mobility data set: 300 cabs in SanFrancisco.
Initially infected node: one cab is randomly chosen.
Running period: 12 days.
Wireless transmission range: Bluetooth (10m), WiFi (100m)
Vulnerability ratio: 10% - 80%
15 / 25
![Page 34: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/34.jpg)
Experimental Evaluation: Results
The size of botnet with two different initially infected nodes andκ = 80%.
0 4 8 12
50
100
150
200
Days
Siz
e of
Bot
net
WiFi
Bluetooth
16 / 25
![Page 35: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/35.jpg)
Experimental Evaluation: Results
The size of botnet with different vulnerability ratios κ and WiFi.
0 12 24 36 48 60
20
40
60
80
100
Hours
Siz
e of
Bot
net
0.4
0.6
0.8
0.1
16 / 25
![Page 36: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/36.jpg)
Summary: EPFL Data Set
For different setups, we always observe the quadratical increase ofthe botnet size!
Different vulnerability ratios
Different transmission ranges
Different initially infected nodes
Reason: Cab movements during 12 Days
sufficient mobility in San Francisco area.
mobility radius α is large.
17 / 25
![Page 37: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/37.jpg)
Summary: EPFL Data Set
For different setups, we always observe the quadratical increase ofthe botnet size!
Different vulnerability ratios
Different transmission ranges
Different initially infected nodes
Reason: Cab movements during 12 Days
sufficient mobility in San Francisco area.
mobility radius α is large.
17 / 25
![Page 38: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/38.jpg)
Experiments with Limited Mobility
UDelModels: a tool to generate realistic mobility traces.Map: 2000 nodes in 2km×2km downtown Chicago, κ=60%, r= 10m (bluetooth),Mobility radius α=10, 100, 500, 1000m.
18 / 25
![Page 39: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/39.jpg)
Experimental Results
The botnet size with different mobility radius α.
0 4 8 12 16 20 24
200
400
600
800
Hours
Siz
e of
Bot
net
10m
1km
100m500m
19 / 25
![Page 40: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/40.jpg)
The Impact of Botnet Attacks
Quadratic growth: A botnet can become larger and larger
Launching attacks targeting a mobile service. [Traynor ’09]
Infected nodes flood service requests.
20 / 25
![Page 41: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/41.jpg)
The Impact of Botnet Attacks
Quadratic growth: A botnet can become larger and larger
Launching attacks targeting a mobile service. [Traynor ’09]
Infected nodes flood service requests.
Question: If a botnet starts to propagate at time 0, how long thebotnet is able to launch an attack to take down a service?
20 / 25
![Page 42: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/42.jpg)
The Impact of Botnet Attacks
Performance Metric: Last Chipper Time
time
service quality
requirement
0 last chipper time
The last time that a required ratio (σ < 1) of mobile servicerequests can still be processed on time under the botnet attack,
Tl = sup{t ≥ 0 : P(Dp < d) > σ}.
20 / 25
![Page 43: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/43.jpg)
The Impact of Botnet Attacks
Theorem: Last chipper time decreases on the order of 1/√
B
network bandwidth B
last chipper time
B
decreasing on the
order of 1 /
20 / 25
![Page 44: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/44.jpg)
The Impact of Botnet Attacks
Theorem: Last chipper time decreases on the order of 1/√
B
network bandwidth B
last chipper time
B
decreasing on the
order of 1 /
Increasing network bandwidth:
improves network performance
a botnet can propagate for a shorter time to disrupt a service.
less time to detect and respond the attack!
20 / 25
![Page 45: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/45.jpg)
The Impact of Botnet Attacks
Theorem: Last chipper time decreases on the order of 1/√
B
network bandwidth B
last chipper time
B
decreasing on the
order of 1 /
Example: LTE→LTE Advanced (10 times bandwidth increase).Last chipper time becomes 1/
√10 ≈ 1/3 of the time in LTE.
20 / 25
![Page 46: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/46.jpg)
Experimental Evaluation
Experimental setups
...
...
...
service provider the network
service
processed
simulation server
requests
results
...
The Network: 2km×2km downtown Chicago, 25 APs
21 / 25
![Page 47: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/47.jpg)
Experimental Evaluation
Experimental setups
...
...
...
service provider the network
service
processed
simulation server
requests
results
...
Service provider: small-scale
7 computers over Storm framework (real-time distributedprocessing).
Service quality requirement: 90% on time.
Service timing requirement: 2 seconds.
21 / 25
![Page 48: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/48.jpg)
Experimental Results
The last chipper time with different mobility radius, κ=60%.
10 18 26 34 42 506
8
10
12
14
16
18
Bandwidth [MHz]
Last
Chi
pper
Tim
e [H
ours
]
1km
100m
500m
22 / 25
![Page 49: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/49.jpg)
Experimental Results
The last chipper time with different mobility radius, κ=60%.
10 18 26 34 42 506
8
10
12
14
16
18
Bandwidth [MHz]
Last
Chi
pper
Tim
e [H
ours
]
1km
100m
500m
Last chipper time decreases on the order of 1/√
B
Increasing B increases the risk of service being disrupted.
22 / 25
![Page 50: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/50.jpg)
Outline
1 Motivation
2 Preliminaries
3 Results
4 Conclusion
23 / 25
![Page 51: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/51.jpg)
Conclusion
1 We investigated how mobile botnets evolve via proximityinfection and their impacts.
2 We found mobility can be a key to the size of a mobilebotnet.
Sufficient mobility → the size increases quadratically over time.Insufficient mobility → the size is bounded by a constant.
3 We defined the metric of last chipper time that offersquantitative risk assessment on potential denial-of-serviceimpacts of botnet attacks in mobile networks.
24 / 25
![Page 52: How Can Botnets Cause Storms? Understanding the Evolution ...](https://reader033.fdocuments.in/reader033/viewer/2022051406/627d2d65c27bcc5b81566084/html5/thumbnails/52.jpg)
Thank you!
25 / 25