Embedded Device Flaws Underscore Infrastructure Vulnerability

The embedded systems used to monitor servers have vulnerabilities that could be used to possibly compromise the systems, researchers say.

The embedded systems used by administrators to monitor servers, known as baseboard management controllers (BMCs), are poorly coded and have numerous vulnerabilities that could be used to disrupt and potentially take control of a system, according to a growing number of research reports.

Earlier this month, researchers from vulnerability management firm Rapid7 found seven security flaws in the Web implementation of SuperMicro's Intelligent Platform Management Interface (IPMI) that put tens of thousands of Internet-accessible servers at risk. The vulnerabilities, many of which have been patched, will likely remain in production devices because embedded devices are not often patched.

"This stuff is typical for embedded device development," said HD Moore, chief research officer for Rapid7. "The same bugs we find here are going to apply to every device on the planet."

Small, but fully functional computers are often embedded in devices and appliances. Such embedded systems can be hacked like any other computer: Researchers have focused on attacking automotive systems, home routers, industrial control systems and a wide variety of consumer appliances. In an early investigation into one model of home router, for example, Rapid7 researchers found 30 vulnerabilities using an automated 10-minute scan.

BMCs are also fully functioning computers, typically running on operating systems such as some flavor of Linux. The devices communicate through IPMI, which has some significant vulnerabilities. In an extensive survey of IPMI vulnerabilities, security researcher Dan Farmer found that flaws in the protocol, the vendor's implementation of the protocol and how end users configure the systems cause Internet-connected servers with BMCs to be vulnerable.

Access to IPMI is almost equivalent to having physical access to the server, according to Oded Horovitz, CEO of PrivateCore, a data protection firm.

"If you have a malicious admin installing malware, then the same can happen in IPMI," he said. "Don't trust IPMI at this point to provide adequate security."

Companies should put IPMI servers behind firewalls and not allow them to directly access the Internet, said Rapid7's Moore. Hosted and cloud services should also make sure that the servers do not have direct access to each other.

Rapid7 regularly scans the Internet and found that 35,000 IPMI-enabled servers allowed a direct connection at the time of the report. The number has since declined to 28,000, Moore said. The issues will be slow to patch, especially because most companies are loath to shut down a server unless there is proof of a security serious flaw, he said.

"My guess is that most of these devices are still vulnerable," Moore said.

Robert Lemos

Robert Lemos

Robert Lemos is an award-winning freelance journalist who has covered information security, cybercrime and technology's impact on society for almost two decades. A former research engineer, he's...