PROJECT OVERVIEW
PROBLEM
Over the past few years, IoT has become one of the most adapted technologies of the 21st century which has enabled us to connect everyday objects to the internet via embedded devices, ensuring seamless communication between people, processes, and things. A new forecast from IDC (International Data Corporation) estimates that there will be 41.6 billion connected IoT devices, or “things,” generating 79.4 ZB of data in 2025. Yet, many of these IoT devices are fundamentally insecure due to their computational and resource constraints, which makes it difficult to implement strong security mechanisms in them. This often makes IoTs an easy target for illegitimate attack and to gain control of. The danger exposed by infested IoT devices threatens the complete Internet ecosystem.
APPROACH
Attack and anomaly detection in IoT has become a major concern now with the increased use of IoT infrastructure in every domain. I aim to inculcate and compare performance metrics of various Machine learning algorithms to predict anomalies and attack traffics accurately.
What is an IoT?
IoT involves adding internet connectivity to a system of interrelated computing devices, mechanical and digital machines, objects, animals and/or people. These are provided with a unique identifier and the ability to automatically transfer data over a network.
What is not an IoT?
The term IoT is mainly used for devices that wouldn't usually be expected to have an internet connection, and that can communicate with the network independently of human action. Hence a PC or a smartphone isn't considered as an IoT, whereas a smartwatch, a fitness band or other wearable device might be counted as an IoT device.
What is Machine Learning?
Machine learning is a method of data analysis that automates analytical model building. It is a branch of artificial intelligence based on the idea that systems can learn from data, identify patterns and make decisions with minimal human intervention. As models are exposed to new data, they learn from previous computations to produce reliable, repeatable decisions and results.
Why integrate ML and IoT?
Anomaly detection has been used in network intrusion detection systems (NIDS) for detecting unwanted behavior in networks for a long time. However, there has been little work tailoring anomaly detection specifically for IoT networks as IoT traffic is different from other types of network behavior since IoT traffic access only specific set of points/nodes and not a large variety of web servers, as their functionality is very specific to that particular application. Fortunately, with IoT, since the functionality of each device is very limited, it is much harder to sneak in malicious requests and much easier to establish a finite set of rules to determine normal and anomalous behavior. Furthermore, the utility of the huge data generated by the IoT is better utilized with the Machine Learning techniques which enable the IoT systems to make informed and intelligent decisions.
Project Timeline
March 22nd Model Comparison & Conclusion Completed
March 3rd - March 21st Data classification & Modeling Completed
March 2nd Mid-Term & Proposal update Completed
February 15th - February 29th Data Pre-Processing Completed
February 8th to February 14th Exploratory Data Analysis (EDA) Completed
February 1st - February 7th Dataset Collection Completed
January 31st 2020 Project Proposal - I Completed
CONCLUSION
Random Forest technique predicted attacks accurately compared to other approaches with very less number of mis-classifications for two of the traffic types (Dos & Normal)
Relying on these estimations, it can be concluded that RF is the best technique for this particular study.
.png)
REFERENCES
Rohan Doshi, Noah Apthorpe, Nick Feamster. “Machine learning DDoS detection for consumer IoT” In: 2018 IEEE Security and Privacy Workshops (SPW). May 2018. DOI : 10.1109/SPW.2018.00013.
Fatima Hussain, Rasheed Hussain, Syed Ali Hassan, and Ekram Hossain. “Machine Learning in IoT Security: Current Solutions and Future Challenges” Available at : https://arxiv.org/pdf/1904.05735.pdf