First Public Demo of Data Breach via IoT Hack Comes to RSAC

RSA CONFERENCE 2018 – San Francisco – Many security professionals acknowledge that Internet of Things (IoT) devices have the potential to be an avenue into their enterprise networks — but for most, breach-by-refrigerator or DDoS-by-coffeepot is a theoretical flight of fancy and not a genuine threat. That might change Thursday, when researchers will present here the first public demonstration of an IoT hack resulting in a breach of personally identifiable information.  

The vice president of research, M. Carlton, and chief technology officer, Stephen Ridley, of IoT security company Senrio will present "Lateral Attacks between Connected Devices in Action" on the RSA Sandbox's IoT stage Thursday. 

"'Chained attacks on IoT security' — it's only been uttered as this platitude," says Ridley, "but have you actually seen a camera get popped" and used to compromise other systems?

"We all know IoT is vulnerable," says Carlton. "We don't all know what the impact of one vulnerable IoT device in an enterprise can be. ... It is a profound impact."

This particular attack can also be a danger to organizations with good security measures in place. In the demo, the IoT device need not be directly connected to the target network device. It doesn't require sophisticated hacking skills — Metasploit tools or the Linux command line will suffice. 

And the attacker never interacts with the endpoint, where most enterprises invest most of their security protections. As the Senrio team puts it, by staying away from the endpoint, the attacker doesn't need to come up against Carbon Black or CrowdStrike.

"This could be done on a company that has spent millions on security," says Ridley. "If I was a bad guy, I'd be doing nothing but IoT. Straight up." 

The attack begins with an exploit of a surveillance camera via the Devil's Ivy vulnerability — a remote code execution vulnerability in an open source gSOAP library that was discovered by the Senrio team last summer. A patch for the vulnerability already exists but was not applied to this camera model — and that's not unusual. 

"In the IoT world, most patches do not get applied," says Ridley. That's due in part to the complexity of the IoT supply chain and the fact that most organizations do not know what IoT devices are connected to their network in the first place.

Once the camera is compromised, the attackers then have a bird's-eye view of an employee at his workstation and the items on his desk — which include a router and a network access server (NAS). The attackers can then watch the user's keystrokes when logging in to the NAS.

The attackers then send a request to the router to obtain its exact model number (so it can retrieve the proper exploit for it), which the router obligingly sends. 

The exploited router sends the attackers encrypted text containing the end user's concatenated username and password. Then, using Rainbow Tables, the attackers can reverse the hash function and determine the administrator credentials for the router. (In this case, username: admin and password: admin.) 

With those credentials in hand, the attackers have full access to the router, which allows them to, among other things, change network settings — which thereby lets them open a secure SSH communication to the NAS and enjoy privileged access to all of the files it contains. 

Owning the NAS, the attackers can thus access all manner of sensitive data, from financial records to personally identifiable information. They copy it and exfiltrate it back through the router, through the video camera, and back home to the attackers.  

How can enterprises defend against attacks like these? Carlton takes a deep breath. 

"First, find what [IoT] devices are on your network," she says. "Then we'll talk."    

 Related content:

Join Dark Reading LIVE for an intensive Security Pro Summit at Interop ITX and learn from the industry’s most knowledgeable IT security experts. Check out the agenda here. Register with Promo Code DR200 and save $200.