The explosion of network-enabled devices embodied in the Internet of Things (IoT) promises amazing advances in convenience, efficiency and even security. But every promising new device generates new seams and potential opportunities for hackers to worm their way into networks and exploit network weaknesses.
Figuring out which IoT devices are safe – and which aren’t – and how to safely leverage the promise of that technology will require looking beyond traditional supply chain and organizational boundaries and developing new ways to approve, monitor and review products that until recently weren’t even on the radar of information security officials.
Conventional product definitions have fundamentally changed, said Dean Souleles, chief technology officer at the National Counterintelligence & Security Center, part of the Office of the Director of National Intelligence. To illustrate his point, he held up a light bulb during the recent Institute for Critical Infrastructure Technology Forum, noting that looks can be deceiving.
“This is not a light bulb,” he said. “It does produce light – it has an LED in it. But what controls the LED is a microcircuit at the base.” That microcircuit is controlled by software code that can be accessed and executed, via a local WiFi network, by another device. In the wrong hands – and without the proper controls –that light bulb becomes a medium through which bad actors can access and exploit a network and any system or device connected to it.
“When my light bulb may be listening to me,” Soules says, “we have a problem.”
What’s On Your Network?
Indeed, the whole government has a problem. Asset management software deployed across 70 federal agencies under the Department of Homeland Security’s Continuous Diagnostics and Mitigation program has uncovered the surprising extent of unknown software and systems connected to government networks: At least 44 percent more assets are on agency networks than were previously known, according to CDM program documents, and in some agencies, that number exceeded 200 percent.
IoT will only make such problems worse, because IT leaders rarely have a comprehensive of the many products acquired and installed in their buildings and campuses. It’s hard enough to keep track of computers, laptops, printers and phones. Reining in facilities managers who may not be fully aware of cyber concerns or properly equipped to make informed decisions is a whole different story.
“You have to have a different way of thinking about the internet of things and security,” Souleles said. “When you think about infrastructure today, you have to think beyond your servers and devices.”
Securing IoT is now a supply chain risk management issue, greatly expanding the definition of what constitutes the IT supply chain. “That risk management has got to [focus on] software risk management,” Souleles said. “You have to begin with the fact that software now includes your light bulbs. It’s a different way of thinking than we have had before. And things are moving so quickly that we really have to stay on top of this.”
The Intelligence Community and the technology companies that support it may be best equipped to define the necessary best practices and procedures for ensuring a safe and secure supply chain. Chris Turner, solutions architect at General Dynamics Information Technology, said the supply chain attack surface is huge, and that the risks among technology products can be huge, as well – if left unattended.
Indeed, Jon Boyens, a senior advisor for information security at the National Institute of Standards and Technology (NIST), says as much as 80 percent of cyber breaches originate in the supply chain, citing a 2015 study by the Sans Institute.
Boyens cites two notorious examples of supply chain failures: In the first, supplier-provided keyboard software gave hackers access to the personal data of 600,000 Samsung Galaxy smartphones; in the second, supplier-provided advertising software let attackers snoop on browser traffic on Lenovo computers. Counterfeit products, devices compromised in transit and component-level vulnerabilities are other supply chain risks that can lead to devastating consequences.
Maintaining sufficient controls to minimize risk and maximize transparency requires close relationships with vendors, clear understanding of the risks involved and strict adherence to procedure. Organizations should be able to identify their lower-tier sub-contractors as well as the extent to which their suppliers retain access to internal systems and technology.
For companies that routinely support highly classified programs, this kind of diligence is routine. But many are not sufficiently experienced to be so well equipped, says GDIT’s Taylor, where supply chain risk management is considered a core competency.
“How do you protect the supply chain when everything comes from overseas?” Taylor asks rhetorically. “You can’t know everything. But you can minimize risk. Experienced government contractors know how to do this: We know how to watch every single component.”
That’s not just hyperbole. For the most classified military systems, source materials may be tracked all the way back to where ore was mined from the Earth. Technology components must be understood in all their infinite detail, including subcomponents, source code and embedded firmware. Minimizing the number of suppliers involved and the instances in which products change hands is one way to minimize risks, he said.
Certified Cyber Safe
Making it easier to secure that supply chain and the software that drives IoT-connected devices is what’s behind a two-year-old standards effort at Underwriters Laboratories (UL), the independent safety and testing organization. UL has worked with the American National Standards Institute (ANSI) and the Standards Council of Canada (SCC) to develop a series of security standards that can be applied to IoT devices from lights, sensors and medical devices to access and industrial controls.
The first of these standards will be published in July 2017 and a few products have already been tested against draft versions of the initial standard, UL 2900-1, Software Cybersecurity for Network-Connectable Products, according to Ken Modeste, leader of cybersecurity services at UL. The standard covers access controls, authentication, encryption, remote communication and required penetration and malware testing and code analysis.
Now it’s up to users, manufacturers and regulators – the market – to either buy into the UL standard or develop an alternative.
Mike Buchwald, a career attorney in the Department of Justice’s National Security Division, believes the federal government can help drive that process. “As we look to connected devices, the government can have a lot of say in how those devices should be secured,” he said at the ICIT Forum. As one of the world’s biggest consumers, he argues, the government should leverage the power of its purse “to change the market place to get people to think about security.”
Whether the government has that kind of market power – or needs to add legislative or regulatory muscle to the process is still unclear. The United States may be the world’s single largest buyer of nuclear-powered submarines or aircraft carriers, but its consumption of commercial technology is small when compared to global markets, especially so when considering the global scale of IoT connections.
Steven Walker, acting director of the Defense Advanced Research Projects Agency (DARPA), believes the government’s role could be to encourage industry.
“What if a company that produces a software product receives something equivalent to a Good Housekeeping Seal of Approval for producing a secure product?” he said in June at the AFCEA Defensive Cyber Operations Conference in Baltimore. “What if customers were made aware of unsecure products and the companies that made them? I’m pretty sure customers would buy the more secure products – in today’s world especially.”
How the government might get involved is unclear, but there are already proven models in place in which federal agencies took an active role in encouraging industry standards for measurement and performance of consumer products.
“Philosophically, I’m opposed to government overregulating any industry,” he said. “In my view, overregulation stifles innovation – and invention. But the government does impose some regulation to keep Americans safe: Think of crash tests for automobiles. So should the government think about the equivalent of a crash test for cyber before a product – software or hardware – is put out on the Internet? I don’t know. I’m just asking the question.”
Among those asking the same question include Rep. Jim Langevin (D-R.I.), an early and outspoken proponent of cybersecurity legislation. “We need to ensure we approach the security of the Internet of Things with the techniques that have been successful with the smart phone and desktop computers: The policies of automatic patching, authentication and encryption that have worked in those domains need to be extended to all devices that are connected to the Internet,” he told the ICIT Forum. “I believe the government can act as a convener to work with private industry in this space.”
What might that look like? “Standard labeling for connected devices: Something akin to a nutritional label, if you will, for IoT,” he said.
Langevin agrees that “the pull of federal procurement dollars” can be an incentive in some cases to get the private sector to buy in to that approach.
But the key to rapid advancement in this area will be getting the public and private sectors to work together and buy into the idea that security is not the sole purview of a manufacturer or a customer or someone in IT, but rather everyone involved in the entire process, from product design and manufacture through software development, supply chain management and long-term system maintenance.
As IoT expands the overall attack surface, it’s up to everyone to manage the risks.
“Pure technological solutions will never achieve impenetrable security,” says Langevin. “It’s just not possible. And pure policy solutions can never keep up with technology.”