Uma heurística para ofuscação de tráfego em internet das coisas

Abstract

The use of smart technologies is more present in population daily lives thank to the advance of the Internet of Things (IoT). Smart Devices capture user’s iteration with the environment, habits, localization, health status, and other information and share these by internet to be processed by remote servers or consumed by several services or applications, that way providing customized user experience. Researches shows that intruders can use patterns present in traffic attributes, like packets size, send intervals and sent packet number to infer user information, that way invading his privacy. In current literature, many methods were proposed to protect users against attacks based on traffic analysis. The review of those works shows that, in general, redundant data are added to the network, what can cause delays and overloads. In this context, its necessary a trade-off analysis before applying obfuscation methods in a IoT network. This project propose an obfuscation approach based in packet padding to reduce attacks based on traffic analysis that use packet lengths to infer user’s information. To balance the trade-off between performance e privacy, the method use a padding heuristic to minimize the number of bytes inserted in the packets. In addition, the results show an improvement up to 90% in the tradeoff between the reduction of attacker success rate and the number of bytes inserted in the packets.