The Internet of Things is bringing every aspect of our lives online. Phones, watches, printers, thermostats, lightbulbs, cameras, and refrigerators are only a handful of devices connecting to home and enterprise networks. This web of products is seemingly intended to make everyday tasks more convenient; unfortunately, their weak security gives attackers an easy route in.
"[The IoT] is still a computer on a network, but it's different," says Joseph Carson, chief security scientist with Thycotic. Unlike traditional PCs, the functionality for IoT devices is very specific; further, they're designed to be inexpensive and simple to deploy. As more employees bring devices into the workplace and connect them to Wi-Fi, the challenge to protect them escalates.
Enterprise devices not historically connected to the Internet are now part of the IoT, complicating the issue, adds Deral Heiland, IoT research lead at Rapid7. He points to multi-functional printers, which he says have long been a corporate security risk. Modern printers can control myriad functions, send data over the Internet, or print remotely via the cloud.
"One of the big things I run into at a lot of organizations is, 'What really is the IoT?'" he says. "Things that weren't on the IoT a decade ago, which have always been in the environment, have morphed into IoT technology." As a result, many businesses don't understand the full breadth of devices putting them at risk.
Routers, printers, and IP cameras are among the most prominently discussed devices in corporate IoT security. Cybercriminals are studying the IoT attack surface, figuring out what works and doesn't work, and how they can profit from vulnerabilities in connected devices. A recent Trend Micro report sheds light on how attackers profit from the IoT: many sell access to hacked IoT devices built into botnets; others extort owners of connected industrial equipment.
In particular, security experts point to the Mirai botnet as a turning point for connected device security. Mirai and its variants "seem to be the big one these days," says Jon Clay, Trend Micro's director of global threat communications. The botnet has "stifled creativity" in the underground for this type of malware: it's open-source and free, so attackers don't have to work very hard.
"The attack surface is growing incrementally," he says. "There are so many new devices coming online." Criminals are narrowing their focus on IoT, evolving from ransomware or point-of-sale malware to specifically targeting connected devices.
Compounding the danger of IoT threats is the rise of nation-state attackers, who are targeting firmware at scale or leveraging connected devices in DDoS attacks. They don't have to attack a major entity in order to have far-reaching effects, either: as NotPetya demonstrated, a nation-state actor could target one single component supplier to have devastating consequences.
Organizations' attitude toward IoT security is similar to their approach to smartphones several years back, Heiland says. Now, they're in the early stages of how they'll improve their business model and put together processes to stay secure. At the same time, standards and regulations are emerging to inform manufacturers how to build security into these devices from the start.
Where Businesses and Manufacturers Fall Short
A combination of poor device security and higher interest among attackers is driving businesses to pay more attention to the IoT. "The attack surface they're responsible for has grown so immensely," says Mike Janke, CEO of DataTribe, where a group of advisory CISOs uses the term "shadow IoT" to refer to the smartwatches, headphones, and tablets appearing on networks.
"That's a big pain because [the CISOs] are ultimately responsible," he continues, noting most don't have the budget, people, or resources to combat the problem. "It's very frustrating."
Many companies continue to struggle with patch management efforts, adds Clay, which adds to the challenge as IoT device manufacturers typically require users to apply updates. "A lot of these devices aren't traditional PCs," he explains. "Even though they have operating systems and applications inside, they aren't treated like a server or a PC is in an organization."
Carson advises organizations to consider the function of IoT devices before permitting them on a corporate network. Is it a data collector or aggregator? Can the rest of the network be accessed through the device? Does it introduce new threats? Who owns the device; can they view or download data? He suggests personal devices be required to access a guest network.
(Continued on Next Page)
Experts agree manufacturers should be playing a bigger role in how their products are secured. Most, says Carson, prioritize ease of use, ease of deployment, and low cost over device security.
There are two categories of IoT device manufacturers, Heiland explains. The first, he says, are "white label" manufacturers selling blank devices other companies can put their names on. This often becomes a bigger security issues, he adds, because it's difficult to trace a product back to its manufacturer if vulnerabilities are found. Rapid7 researchers faced this precise vendor visibility challenge when they investigated bugs in children's GPS-enabled smartwatches.
The other category includes brand-name tech companies, which over the past few years have expanded their security involvement and let vulnerabilities be reported. These companies, says Heiland, are better prepared to adopt emerging guidelines and requirements because they have sufficient resources. White-label manufacturers, on the other hand, may not be ready.
"We would argue people should probably getting more security audits throughout the development process," says Shaunak Mirani, researcher with Independent Security Evaluators (ISE), which recently published updated research on its analysis of routers and network-attached storage (NAS) devices. What they found are high-severity vulnerabilities that enable root compromise of a router or NAS from an attacker who has network-level access to the device.
Pivoting activity, says Mirani, is the major threat he sees with IoT devices. "One IoT device by itself may not have a huge risk factor, but when you start chaining things together, and get one device to talk to another, and dig deeper into the network … that's the real threat."
Overall, Clay says, manufacturers are starting to realize security is a critical component of device development, manufacturing, and release; however, "they're slow on the uptake of that." A lack of activity is pushing federal and state regulators to create guidelines and rules.
Regulations and Standards and Ratings, Oh My!
One of these measures is California's IoT Device Security Act (SB-327), which is poised for implementation on Jan. 1, 2020. The bill requires IoT device manufacturers to equip connected products with "reasonable security feature or features" appropriate to the nature or function of the device; appropriate to the data it may collect, hold, or transmit; and designed to protect the device and its data from unauthorized access, destruction, use, modification, or disclosure.
If a connected device has means for authentication outside a local area network, the bill states, it will be considered a security feature if the preprogrammed password is unique to each device, and if the device requires its user to create new credentials before gaining initial access. Even with these specifications, the bill's language is ambiguous, notes Harley Geiger, director of public policy for Rapid7, who points out the law was replicated in Oregon this summer. It doesn't say how "reasonable security" varies among devices, or how the bill will be enforced.
"Whether the companies will meet these obligations will depend on how realistic they are," Geiger continues. Some draft regulations would require fixing every vulnerability to achieve a standard of reasonable security, which he says isn't consistent with risk management practices.
California's bill isn't the only measure aiming to improve IoT security. Some, like this one, are requirements. Others, like NIST's "Core Cybersecurity Feature Baseline for Securable IoT Devices" released this summer, are more like "strongly worded guidelines," Geiger adds.
Indeed, NIST's core baseline details "voluntary recommended cybersecurity features" to build into network-capable devices. These include device identification for connecting to a network; device configuration to provide an option for changing software and firmware; data protection for information stored and sent over the network; limited access to local and network interfaces; updatable software and firmware; and accessible cybersecurity event logging.
Across the pond, the UK published its own Code of Practice for Consumer IoT Security, which contains 13 outcome-focused guidelines detailing best IoT security practices for manufacturers and other industry bodies. It also emphasizes default passwords in its first guideline: "All IoT device passwords shall be unique and not resettable to any universal factory default value."
Its second guideline advises vulnerability disclosure policies and suggests businesses provide a public point-of-contact for security researchers to report issues. The Code of Practice also encourages timely and secure software updates, secure credential storage (hard-coded credentials are discouraged), and encrypted transmission of security-sensitive information.
(Continued on Next Page)
"The focus from the EU is a good one, it is important, and we have to take that into consideration, especially as it comes to end nodes," says Dr. Joerg Borchert, president and chairman of Trusted Computing Group (TCG). Privacy protection is a higher priority in Europe, he adds, and it will be a critical part of the conversation around IoT security. As TCG focuses on IoT security, the organization has been working closely with several governments and standards organizations.
"We try to understand what kind of best practices can be utilized and also, it is important for an industry standard to harmonize as much as possible across different geographies and different countries," says Borchert.
UL's IoT Security Rating is another industry measure geared toward manufacturers. Its evaluation process considers critical security aspects of connected products against common attack strategies and known IoT vulnerabilities to create a "security baseline" for consumers.
The driver for UL's rating was to "incentivize manufacturers to build security into their products," says director of security and technology Andrew Jamieson, who anticipates a consumer demand for a minimal security baseline. Adding security will increase cost, he adds, but advertising secure devices beside unsecured ones may encourage people to pay more.
"One of the issues we have with security is it's a commercial problem as much as it is a technical problem," Jamieson explains. He compares the IoT security rating to energy ratings on tools and appliances: because consumers understand why the cost is higher, they're likely to choose a more energy-efficient product. Security ratings will vary between low-risk products, like a connected lightbulb, and high-risk products such as wireless and IP-connected cameras.
IoT Security Startups Bring New Ideas, Capabilities
In addition to providing a gateway into target networks, insecure IoT devices can grant access to a wealth of personal data. Potential exposure of this information is another factor driving private and public sector organizations to pay closer attention to how devices are secured.
"When you think about the amount of data and everything being connected, whether it's at home, on your body, how you drive to work, the threat vector is just growing in magnitudes that you can barely comprehend today," says Gregg Smith, CEO of startup Attila Security. The company launched in 2018 to protect endpoints using a software-defined perimeter.
Attila's tech comes from the NSA, Smith explains. Its initial use case was to provide traveling executives secure connectivity back into government networks. Over time, the company has expanded its use cases to organizations across governments and industries. Now it enables secure IoT deployments, sensitive communications, and secure remote network access. Channels connect devices to one another, enabling IoT device security at a larger scale.
Securing communications across devices is "solving a problem that IoT is creating, but it's not attacking the underlying problem," says Janke. Going down to a deeper level is ReFirm Labs, another IoT security startup specifically focused on the analysis and vetting of IoT firmware.
Firmware, an appealing target given its higher level of access and privilege on a device, is a growing concern in the IoT security industry because it's commonly unprotected. ReFirm's Centrifuge Platform validates and monitors the security of firmware running billions of IoT devices and connected enterprise machines.
"It takes just one firmware weakness for bad actors to gain access to an IoT device and then use that attack surface to compromise the integrity of an entire network," says cofounder Terry Dunlap. These attacks often aren't advanced or complicated to perform; intruders can simply take advantage of default usernames and passwords, which come with so many IoT products.
Where We're Headed
In the future, we'll start to see greater monetization of IoT devices and criminals targeting medical devices, robot assemblies, and industrial control systems, Clay predicts. As new devices come online and organizations automate, we'll continue to see new IoT-focused attacks.
Carson calls on industry organizations to share data across verticals, which he believes can help everyone better prepare for IoT attacks. "Sometimes a lot of lessons can be learned by having cross-industry experience," he notes. "We need to talk more about the successes and share more about the lessons learned."
It's "highly likely" we'll continue to see more actions from state and federal agencies to address IoT security, Geiger anticipates, though he believes states' progress will move faster. While major tech organizations like Amazon and Microsoft are taking regulation seriously, more will need to be done to bring manufacturers of all levels on board.