ThingPot: an interactive Internet-of-Things honeypot

TL;DR

This paper introduces ThingPot, an innovative IoT honeypot that simulates entire IoT platforms to study attacker behavior, revealing five attack types over 1.5 months of deployment.

Contribution

It presents the first honeypot focusing on whole IoT platforms, enabling detailed analysis of attack strategies against IoT devices.

Findings

Identified five attack types and vectors against IoT devices

Deployed for 1.5 months, capturing real attacker behavior

Open-sourced the honeypot for community use

Abstract

The Mirai Distributed Denial-of-Service (DDoS) attack exploited security vulnerabilities of Internet-of-Things (IoT) devices and thereby clearly signalled that attackers have IoT on their radar. Securing IoT is therefore imperative, but in order to do so it is crucial to understand the strategies of such attackers. For that purpose, in this paper, a novel IoT honeypot called ThingPot is proposed and deployed. Honeypot technology mimics devices that might be exploited by attackers and logs their behavior to detect and analyze the used attack vectors. ThingPot is the first of its kind, since it focuses not only on the IoT application protocols themselves, but on the whole IoT platform. A Proof-of-Concept is implemented with XMPP and a REST API, to mimic a Philips Hue smart lighting system. ThingPot has been deployed for 1.5 months and through the captured data we have found five types of attacks and attack vectors against smart devices. The ThingPot source code is made available as open source.