E-Locus - Institutional Repository of the University of Crete

Home Search

Results - Details

Search command : Author="Παπαδοπούλη" And Author="Μαρία"

Current Record: 10 of 45

[Add to Basket]

Identifier

000428738

Title

Smart home devices activity fingerprinting from their network traffic and the role of virtual assistant devices

Alternative Title

Ταυτοποίηση έξυπνων οικιακών συσκευών και αναγνώριση των εντολών τους από την διαδικτυακή κίνηση και ο ρόλος των εικονικών βοηθών

Author

Καραμπινάκης, Εμμανουήλ Σ.

Thesis advisor

Μαρκάτος, Ευάγγελος

Reviewer

Παπαδοπούλη, Μαρία
Μαγκούτης, Κωνσταντίνος

Abstract

Smart home devices are tending to become a necessity in our fast-paced life, Internet of Things (IoT) devices have proved to solve many of our everyday life problems of any kind. With the introduction of IoT devices in our homes, security and privacy must be taken more seriously, especially when the nature of those devices is to interact or handle sensitive and personal data. Amazon Echo is a smart speaker that can understand natural language voice commands and complete tasks for the user by communicating with Alexa, a cloud-based virtual assistant where the commands are analyzed and comprehended. The user activates the device with a wake up" word, then he interacts with it with voice commands. Alexa is also capable of connecting with peripheral smart devices at the same WiFi network, thus creating a smart home with interconnected devices under the brain of Alexa. We will demonstrate how an ISP or any other passive network observer with access to a users last-mile network traffic is capable of identifying the types of smart home devices even behind Network Address Translation (NAT). Figure out their state, and how the interconnection of the smart devices with a virtual assistant like Amazon Alexa, in the way that is implemented now, is giving away privacy-sensitive information. We will analyze the security of Amazon Echo and a few different types of smart home devices that are Alexa compatible, and we will describe the process of how a malicious network observer can infer information about the user. Also, we will present and cluster the security of IoT devices, how secure an IoT device can be with the existing implementation and on the other hand, how insecure is a more vulnerable smart device at the point where an attacker can not only extract sensitive information about the user's life but also actively interfere with its usage.

Language

English

Subject

IOT

Network traffic monitorine

Διαδίκτυο των πραγμάτων

Εικονικοί βοηθοί

Μελέτη διαδικτυακής κίνησης

Issue date

2020-03-27

Collection

School/Department--School of Sciences and Engineering--Department of Computer Science--Post-graduate theses