Quantitative Privacy Analysis of Amazon Echo Devices Through Network Traffic Monitoring : A Low-Cost Raspberry Pi Access Point Methodology for Independent IoT Privacy Assessment
avoin
Julkaisu on tekijänoikeussäännösten alainen. Teosta voi lukea ja tulostaa henkilökohtaista käyttöä varten. Käyttö kaupallisiin tarkoituksiin on kielletty.
Lataukset140
Pysyvä osoite
Verkkojulkaisu
DOI
Tiivistelmä
The rising number of Internet of Things (IoT) devices, which include Amazon Echo smart home assistants through voice activation, has created major privacy issues while making it difficult for scientists to study these devices because of expensive research requirements and insufficient analytical tools. This thesis presents a novel Raspberry Pi access point methodology for analyzing IoT device privacy and, independent analysis of Amazon Echo device network behavior.
Three main obstacles confront the existing IoT privacy research: it lacks reproducible results, requires costly commercial instruments that cost over $10,000, and lacks adequate quantitative measurement standards. Through automated packet capture and real-time analysis, machine learning-based device detection, and full data processing, the project develops a low-cost system that leverages a Raspberry Pi 4 as a wireless access point for Amazon Echo network traffic interception and analysis. Because the study approach is totally open-source and only costs $100 instead of $10,000+, researchers can reduce their expenditures by 99% while still achieving better findings
There are significant privacy hazards, according to a review of 168 hours of Echo device traffic data. When the devices are idle, they transmit 151 KB of data every hour, of which 96.5% is sent to Amazon services. The technology produces an ongoing surveillance system by maintaining an ongoing connection at heartbeat intervals of two to three minutes. User behaviour and traffic patterns are strongly correlated, with voice activation resulting in immediate response and prolonged processing after engagement. Network protocol analysis shows that 99.8% of TCP data transactions are encrypted completely, preventing content examination but enabling data transmission monitoring for researchers.
The study revealed three main privacy threats, which consist of continuous monitoring of users when they are not active and the collection of more data than users expect and network traffic analysis to track user activities, and insufficient disclosure because of encrypted communication systems. The regulatory compliance analysis shows that the Amazon Echo does not fulfill GDPR and CCPA requirements, yet the unified data system enables organizations to obtain all user data for business purposes and user information collection.
The system demonstrated 99.8% packet capture accuracy and 95% successful replication during its extended testing periods, which proved its ability to support large-scale multi-device research. The research contributes to IoT privacy studies through its development of user-friendly tools, established performance metrics, and practical guidelines for users, manufacturers, and regulatory bodies. The research provides initial findings for independent IoT device studies, which will lead to the development of privacy-oriented IoT systems.