Today's Internet connects billions of physical devices. These devices are often immature and insecure, and share common vulnerabilities. The predominant form of attacks relies on recent advances in Internet-wide scanning and device discovery. The speed at which (vulnerable) devices can be discovered, and the device monoculture, mean that a single exploit, potentially trivial, can affect millions of devices across brands and continents.

In an attempt to detect and profile the growing threat of autonomous and Internet-scale attacks against the Internet of Things, we revisit honeypots, resources that appear to be legitimate systems. We show that this endeavour was previously limited by a fundamentally flawed generation of honeypots and associated misconceptions.

We show with two one-year-long studies that the display of warning messages has no deterrent effect in an attacked computer system. Previous research assumed that they would measure individual behaviour, but we find that the number of human attackers is orders of magnitude lower than previously assumed.

Turning to the current generation of low- and medium-interaction honeypots, we demonstrate that their architecture is fatally flawed. The use of off-the-shelf libraries to provide the transport layer means that the protocols are implemented subtly differently from the systems being impersonated. We developed a generic technique which can find any such honeypot at Internet scale with just one packet for an established TCP connection.

We then applied our technique and conducted several Internet-wide scans over a one-year period. By logging in to two SSH honeypots and sending specific commands, we not only revealed their configuration and patch status, but also found that many of them were not up to date. As we were the first to knowingly authenticate to honeypots, we provide a detailed legal analysis and an extended ethical justification for our research to show why we did not infringe computer-misuse laws.

Lastly, we present honware, a honeypot framework for rapid implementation and deployment of high-interaction honeypots. Honware automatically processes a standard firmware image and can emulate a wide range of devices without any access to the manufacturers' hardware. We believe that honware is a major contribution towards re-balancing the economics of attackers and defenders by reducing the period in which attackers can exploit vulnerabilities at Internet scale in a world of ubiquitous networked `things'.