Are Older Medical IoT Devices Actually More Secure?

Are Older Medical IoT Devices Actually More Secure?

Cylera Team
By Cylera Team

Are newer medical IoT devices more to blame for healthcare network vulnerabilities?

Legacy medical IoT devices may lack security features, but newer ones built around commodity components can have a whole different set of vulnerabilities that are better understood by attackers. It seems like experts can't agree which is the bigger security problem. 

At first glance, you might think that older devices still integrated into the network are the main weakness in stories of insecure IoT. Some of these devices even pre-date the internet by a considerable length of time. And even if the devices aren’t quite that old, they often lack key capabilities – in particular, remote software updates and configurable password protection – that would help IT staff defend them against modern threats. This presents unique challenges in trying to secure them against remote compromise

That might not be strictly true in regards to the medical field, according to Richard Staynings, chief security strategist for Cylera. He argues that there has been an explosion in the number and variety of medical IoT devices in recent years, and many of those gadgets are at least as insecure as the legacy equipment in the field. In some cases, the older devices might actually be considerably more secure than those of more recent vintage. In particular, those based on dated technology like older versions of electrically erasable programmable read-only memory (EEPROM).

“The older systems were written in EEPROM – you need an EEPROM reader to mess with them, and the codebase is not on the Internet for hackers to look at. You need physical access to the EEPROM to rewrite it,” says Staynings. 

In contrast, the newer devices frequently use software and hardware components that are much more familiar or accessible to potential attackers.

“They’re more ubiquitous in their design and construction – they use [consumer off-the-shelf] operating systems, like Windows embedded, which is still being used believe it or not, and they are much more vulnerable to attack than a legacy system,” Staynings said.

Insecurity in the current generation of medical IoT hardware also carries the potential for ongoing problems, not just immediate ones. While IT assets get replaced rapidly, IoT devices often have much longer replacement cycles.

“Medical devices have the half-life of plutonium,” said Staynings. “They just don’t go away.”

The Debate

Other experts are less sold on Staynings’ characterization of the threat to medical IoT, arguing that the idea that newer devices pose a greater threat than older ones flies in the face of recent efforts to make them safer.

Keith Mularski directs an advisory cybersecurity practice at Ernst and Young, and described Staynings’ assertion as “surprising,” noting that the regulatory landscape for connected medical devices is quickly moving standards in a positive direction.

“The FDA has some pretty stringent guidelines that before devices can go to market, you need to put together threat modeling, so looking at security architecture, vectors, and so on, and then in addition to that the FDA is getting ready to require third-party pen testing in premarket submissions,” said Mularski. “With legacy devices, those premarket submissions aren’t nearly as complete.”

Mularski does concede that some particularly vulnerable old devices are often more isolated on the network by design, in part because they’re more recognizable as vulnerable assets. Windows 95-vintage x-ray machines, for example, are easy to spot as a potential target for a criminal.This underlines a topic most experts can agree on: awareness of the potential security flaws on a given network is key to securing healthcare networks.

“For the most part, I think most of the hospital environments, they do a good job at recognizing that they have these old deices, and the ones that are more vulnerable,” he said.

Greg Murphy, CEO of network visibility and security startup Ordr, said that both Mularski and Staynings have points in their favor.

“Anyone who minimizes the issue of legacy devices needs to walk a mile in the shoes of the biomedical engineering department at a hospital,” he said. “[But] on the flipside, new devices that are being connected to the network have huge vulnerabilities themselves. Many manufacturers themselves don’t know what vulnerabilities their devices have.”

Is There A Clear Answer?

The only real way to address the issues, said Murphy, is at the network level – trying to make everything secure at the device level might be a near-impossibility in many cases, and even getting an accurate picture of every device connected to a network often requires the use of an automated solution.

“This is not a problem of human scale anymore,” he said.

Both Mularski and Staynings concurred on this point as well. Regardless of which devices on a particular network are the most vulnerable, it’s worth remembering that cybercriminals generally aren’t particular about what they compromise, as long as they’re able to gain access. It may be a moot point to continue the argument at this level when risks can come from all angles and access points.

“There may be an attacker that comes across these devices, runs a scan and happens to see [a vulnerability], but we really haven’t seen specific targeting of devices,” said Mularski. “It’s important to make sure that companies that have medical devices are enumerating their network, tracking their devices.”

To learn more about healthcare cybersecurity, visit our resources page. 

Get Updates

Sign up to receive the
latest news from Cylera.

window.lintrk('track', { conversion_id: 14567298 });