Federated Learning for Intrusion Detection

Systems in IoT Devices

Presented By:
Syeda Faiza Ahmed
Roll: 1615

Session: 2021-22
Department of Robotics and Mechatronics Engineering, University of Dhaka
Contents

• IoT
• Attacks on IoT systems
• Intrusion Detection Systems
• Federated Learning
• Case Study: Federated IoT Attack Detection Using
Decentralized Edge Data

2
“Connects the Physical world with the Cyber world”

3
 Figure 1: Growth of IoT devices from 2015 to 2025  [1]

[1]  Al-Bahri, M.; Yankovsky, A.; Borodin, A.; Kirichek, R. Testbed for identify IoT-devices based on digital object architecture. In Internet of
Things, Smart Spaces, and Next Generation Networks and Systems; Springer: Berlin/Heidelberg, Germany, 2018; pp. 129–137 4
 Vulnerable Security of data

Figure 2: Security and privacy of data in IoT [2]

[2] Singh, P. (2018). Cross-layer design for Internet of Things (IoT)-issues and possible solutions. Dep. Syst. Comput. Eng, 1-10.
5
 Attacks on IoT

• Botnet Attacks
• DDoS(Distributed Denial-of-
Service) Attacks
• Firmware Exploitation 
• Data Privacy and Breaches
• Credential Attacks 

Figure 3: Attacks on IoT [3]

[3] Lee, Shu-Hung, et al. “Detection and Prevention of DDoS Attacks on the IoT.” Applied Sciences, vol. 12, no. 23, Dec. 2022, p. 12407.
DOI.org (Crossref), https://doi.org/10.3390/app122312407.  6
 What is an Intrusion Detection System(IDS)?

Figure 4:The Scenario of the IDS applied in the IoT network.[4]

[4] Zhong, M., Zhou, Y., & Chen, G. (2021). Sequential model based intrusion detection system for IoT servers using deep learning methods.
Sensors, 21(4), 1113. 7
Why do we need IDS in IoT Systems?

Figure 5: The Mirai botnet attack that shook the internet in 2016.[5] [6] 

8
 IoT DDoS1 Attack through Mirai Botnet

Figure 06: The Lifecycle of IoT Botnet [7]

1. DDoS=Distributed Denial of Service Attack

[7]  Alzahrani, H., Abulkhair, M., & Alkayal, E. (2020). A multi-class neural network model for rapid detection of IoT botnet attacks. Int. J. Adv. Comp.
Sci. Appl, 11(7), 688-696. 9
 Types of IDS

Figure 07: Intrusion Detection System model.[8]

[8]  Zainel, H., & Koçak, C. (2022). LAN intrusion detection using convolutional neural networks. Applied sciences, 12(13), 6645.
10
Drawbacks of the current IDS systems

• Privacy Concerns
• Communication Overhead
• Single Point of Failure
• Scalability Challenges
• Limited Local Context

11
 What is Federated Learning?
A decentralized form of Machine Learning.

Figure 08:  Simple representation of Federated Learning [9]

[9] Li, Tian, Anit Kumar Sahu, Ameet Talwalkar, and Virginia Smith. "Federated learning: Challenges, methods, and future directions." IEEE Signal
Processing Magazine 37, no. 3 (2020): 50-60. 12
 Machine learning vs Federated learning

Figure 09: Comparison between Traditional ML and FL [10]

[10] Zhou, Xingchen, Ming Xu, Yiming Wu, and Ning Zheng. "Deep model poisoning attack on federated learning." Future Internet 13, no. 3 (2021): 73.
13
 Federated Learning Steps

Figure 10: Steps of Federated Learning [11]

[11] Malandrino, Francesco, and Carla Fabiana Chiasserini. "Federated learning at the network edge: When not all nodes are created equal." IEEE
Communications Magazine 59, no. 7 (2021): 68-73. 14
 Federated Learning Steps

Figure 10: Steps of Federated Learning [11]

[11] Malandrino, Francesco, and Carla Fabiana Chiasserini. "Federated learning at the network edge: When not all nodes are created equal." IEEE
Communications Magazine 59, no. 7 (2021): 68-73. 15
 Federated Learning Steps

Figure 10: Steps of Federated Learning [11]

[11] Malandrino, Francesco, and Carla Fabiana Chiasserini. "Federated learning at the network edge: When not all nodes are created equal." IEEE
Communications Magazine 59, no. 7 (2021): 68-73. 16
 Case Study: Federated IoT attack detection using decentralized
edge data[12]

• Decentralized Federated Learning (FL) for IoT Attack Detection

• Localized Model Training and Aggregation

• Enhanced Attack Detection and Privacy Preservation

[12]Regan, C., Nasajpour, M., Parizi, R. M., Pouriyeh, S., Dehghantanha, A., & Choo, K. K. R. (2022). Federated IoT attack detection using
decentralized edge data. Machine Learning with Applications, 8, 100263. 17
 Case Study: Federated IoT attack detection using decentralized
edge data (Continued)

Figure 11: Overview of proposed architecture [13]

[12]Regan, C., Nasajpour, M., Parizi, R. M., Pouriyeh, S., Dehghantanha, A., & Choo, K. K. R. (2022). Federated IoT attack detection using
decentralized edge data. Machine Learning with Applications, 8, 100263. 18
Summary

• Growth of IoT and Security/Privacy Vulnerabilities

• Role of IDS in IoT Security
• Limitations of Existing IDS Infrastructures
• Federated Learning (FL) for Data Privacy
• Case Study: IDS Framework with Federated Learning

19
Thank You

20
References
[1]  Al-Bahri, M.; Yankovsky, A.; Borodin, A.; Kirichek, R. Testbed for identify IoT-devices based
on digital object architecture. In Internet of Things, Smart Spaces, and Next Generation
Networks and Systems; Springer: Berlin/Heidelberg, Germany, 2018; pp. 129–137
[2] Singh, P. (2018). Cross-layer design for Internet of Things (IoT)-issues and possible solutions.
Dep. Syst. Comput. Eng, 1-10.
[3] Lee, Shu-Hung, et al. “Detection and Prevention of DDoS Attacks on the IoT.” Applied
Sciences, vol. 12, no. 23, Dec. 2022, p. 12407. DOI.org (Crossref),
https://doi.org/10.3390/app122312407. 
[4] Zhong, M., Zhou, Y., & Chen, G. (2021). Sequential model based intrusion detection system
for IoT servers using deep learning methods. Sensors, 21(4), 1113. 
[5] “Someone Is Using Mirai Botnet to Shut Down Internet for an Entire Country.” The Hacker
News, https://thehackernews.com/2016/11/ddos-attack-mirai-botnet.html. Accessed 12 July
2023.
[6] Woolf, Nicky. “DDoS Attack That Disrupted Internet Was Largest of Its Kind in History,
Experts Say.” The Guardian, 26 Oct. 2016. The Guardian,
https://www.theguardian.com/technology/2016/oct/26/ddos-attack-dyn-mirai-botnet.
21
References(Cont.)
[7]  Alzahrani, H., Abulkhair, M., & Alkayal, E. (2020). A multi-class neural network model for rapid
detection of IoT botnet attacks. Int. J. Adv. Comp. Sci. Appl, 11(7), 688-696. 
[8]  Zainel, H., & Koçak, C. (2022). LAN intrusion detection using convolutional neural networks.
Applied sciences, 12(13), 6645. 
[9] Li, Tian, Anit Kumar Sahu, Ameet Talwalkar, and Virginia Smith. "Federated learning:
Challenges, methods, and future directions." IEEE Signal Processing Magazine 37, no. 3 (2020):
50-60. 
[10] Zhou, Xingchen, Ming Xu, Yiming Wu, and Ning Zheng. "Deep model poisoning attack on
federated learning." Future Internet 13, no. 3 (2021): 73. 
[11] Malandrino, Francesco, and Carla Fabiana Chiasserini. "Federated learning at the network
edge: When not all nodes are created equal." IEEE Communications Magazine 59, no. 7 (2021):
68-73
[12]Regan, C., Nasajpour, M., Parizi, R. M., Pouriyeh, S., Dehghantanha, A., & Choo, K. K. R. (2022).
Federated IoT attack detection using decentralized edge data. Machine Learning with
Applications, 8, 100263.
22