When the IoT becomes an agent of the dark side
When parts of the internet were unavailable earlier this month, people took notice. Popular websites like Netflix, Twitter and The New York Times were all but inaccessible thanks to a distributed denial of service attack that used Internet of Things devices to flood Dyn’s managed DNS infrastructure with packet flow bursts 40 to 50 times higher than normal.
In a postmortem released by Dyn, the company said the incident raised important questions surrounding the increasingly common IoT technologies.
The source of the attack was the Mirai malware that infects Linux devices and turns them into bots that it controls for large-scale network attacks. Mirai botnet attacks had been seen by researchers for a few months leading up to last week’s incident, according to Ryan Barnett, a principal security researcher at Akamai who spoke at the Oct. 27 Security of Things forum in Washington, D.C.
The botnet has some unique capabilities, Barnett said. Mirai has up to 10 different attack vectors, it’s globally distributed and it’s segmented — meaning portions of the botnet can attack while others do not. It also can show false IP addresses for where the attacks originated, and the source code is public.
The open source code is “an interesting aspect as well,” Barnett said. “You can confirm certain things you’re seeing, but that also expanded the capabilities of other people taking the code and modifying it.”
Most DDoS attacks rely on a process known as amplification to drive large amounts of traffic to the sites attackers want to crash. Mirai, however, uses a large number of IoT devices to directly attack, he said.
One of the easiest fixes for safeguarding IoT devices against being used in one of these attacks is better credentials. Many IoT devices come with the username “admin” and a common, guessable password. That means any preset password needs to be changed, Barnett said.
This issue must be addressed by manufacturers, but will also require more vigilance on the part of consumers, he said.
“Having a recall would stop the bleeding for right now because it would take things off the shelves,” Barnett said. “But what about all the systems that are already out there?” For currently deployed IoT devices, it’s critical that users get information about turning them off when they’re not in use and upgrading passwords, he said. If the Mirai botnet does infect a device, the password must be reset and the system rebooted to get rid of it.
The lack of accountability for IoT-based security breaches is also an issue, Invincea CEO Anup Ghosh told attendees. Manufacturers aren’t held responsible, and customers may not care if their device is used as a tool in a DDoS attack as long as it doesn’t interfere with what they’re doing, Ghosh said.
“There’s not a built-in incentive for either the manufacturer of the buyer to pay for security,” Ghosh said. “This is an economics- and market-correction issue we’re going to have to deal with.”
It will also become a regulatory issue, according to Stan Lowe, an executive adviser at Booz Allen Hamilton. “I think that’s something the next administration is absolutely going to have to wrestle with.”
