Disaster stories involving the Internet of Things are all the rage. They feature cars (both driven and driverless), the power grid, dams, and tunnel ventilation systems. A particularly vivid and realistic one, near-future fiction published last month in New York Magazine, described a cyberattack on New York that involved hacking of cars, the water system, hospitals, elevators, and the power grid. In these stories, thousands of people die. Chaos ensues. While some of these scenarios overhype the mass destruction, the individual risks are all real. And traditional computer and network security isn’t prepared to deal with them.
Classic information security is a triad: confidentiality, integrity, and availability. You’ll see it called “CIA,” which admittedly is confusing in the context of national security. But basically, the three things I can do with your data are steal it (confidentiality), modify it (integrity), or prevent you from getting it (availability).
So far, Internet threats have largely been about confidentiality. These can be expensive; one survey estimated that data breaches cost an average of $3.8 million each. They can be embarrassing, as in the theft of celebrity photos from Apple’s iCloud in 2014 or the Ashley Madison breach in 2015. They can be damaging, as when the government of North Korea stole tens of thousands of internal documents from Sony or when hackers stole data about 83 million customer accounts from JPMorgan Chase, both in 2014. They can even affect national security, as in the case of the Office of Personnel Management data breach by—presumptively—China in 2015.
On the Internet of Things, integrity and availability threats are much worse than confidentiality threats. It’s one thing if your smart door lock can be eavesdropped upon to know who is home. It’s another thing entirely if it can be hacked to allow a burglar to open the door—or prevent you from opening your door. A hacker who can deny you control of your car, or take over control, is much more dangerous than one who can eavesdrop on your conversations or track your car’s location.
With the advent of the Internet of Things and cyber-physical systems in general, we’ve given the Internet hands and feet: the ability to directly affect the physical world. What used to be attacks against data and information have become attacks against flesh, steel, and concrete.
Today’s threats include hackers crashing airplanes by hacking into computer networks, and remotely disabling cars, either when they’re turned off and parked or while they’re speeding down the highway. We’re worried about manipulated counts from electronic voting machines, frozen water pipes through hacked thermostats, and remote murder through hacked medical devices. The possibilities are pretty literally endless. The Internet of Things will allow for attacks we can’t even imagine.
The increased risks come from three things: software control of systems, interconnections between systems, and automatic or autonomous systems. Let’s look at them in turn:
Software Control. The Internet of Things is a result of everything turning into a computer. This gives us enormous power and flexibility, but it brings insecurities with it as well. As more things come under software control, they become vulnerable to all the attacks we’ve seen against computers. But because many of these things are both inexpensive and long-lasting, many of the patch and update systems that work with computers and smartphones won’t work. Right now, the only way to patch most home routers is to throw them away and buy new ones. And the security that comes from replacing your computer and phone every few years won’t work with your refrigerator and thermostat: on the average, you replace the former every 15 years, and the latter approximately never. A recent Princeton survey found 500,000 insecure devices on the Internet. That number is about to explode.
Interconnections. As these systems become interconnected, vulnerabilities in one lead to attacks against others. Already we’ve seen Gmail accounts compromised through vulnerabilities in Samsung smart refrigerators, hospital IT networks compromised through vulnerabilities in medical devices, and Target Corporation hacked through a vulnerability in its HVAC system. Systems are filled with externalities that affect other systems in unforeseen and potentially harmful ways. What might seem benign to the designers of a particular system becomes harmful when it’s combined with some other system. Vulnerabilities on one system cascade into other systems, and the result is a vulnerability that no one saw coming and no one bears responsibility for fixing. The Internet of Things will make exploitable vulnerabilities much more common. It’s simple mathematics. If 100 systems are all interacting with each other, that’s about 5,000 interactions and 5,000 potential vulnerabilities resulting from those interactions. If 300 systems are all interacting with each other, that’s 45,000 interactions. 1,000 systems: 12.5 million interactions. Most of them will be benign or uninteresting, but some of them will be very damaging.
Autonomy. Increasingly, our computer systems are autonomous. They buy and sell stocks, turn the furnace on and off, regulate electricity flow through the grid, and—in the case of driverless cars—automatically pilot multi-ton vehicles to their destinations. Autonomy is great for all sorts of reasons, but from a security perspective it means that the effects of attacks can take effect immediately, automatically, and ubiquitously. The more we remove humans from the loop, faster attacks can do their damage and the more we lose our ability to rely on actual smarts to notice something is wrong before it’s too late.
We’re building systems that are increasingly powerful, and increasingly useful. The necessary side effect is that they are increasingly dangerous. A single vulnerability forced Chrysler to recall 1.4 million vehicles in 2015. We’re used to computers being attacked at scale—think of the large-scale virus infections from the last decade—but we’re not prepared for this happening to everything else in our world.
Governments are taking notice. Last year, both Director of National Intelligence James Clapper and NSA Director Mike Rogers testified before Congress, warning of these threats. They both believe we’re vulnerable.
This is how it was phrased in the DNI’s 2015 Worldwide Threat Assessment: “Most of the public discussion regarding cyber threats has focused on the confidentiality and availability of information; cyber espionage undermines confidentiality, whereas denial-of-service operations and data-deletion attacks undermine availability. In the future, however, we might also see more cyber operations that will change or manipulate electronic information in order to compromise its integrity (i.e. accuracy and reliability) instead of deleting it or disrupting access to it. Decision-making by senior government officials (civilian and military), corporate executives, investors, or others will be impaired if they cannot trust the information they are receiving.”
The DNI 2016 threat assessment included something similar: “Future cyber operations will almost certainly include an increased emphasis on changing or manipulating data to compromise its integrity (i.e., accuracy and reliability) to affect decision making, reduce trust in systems, or cause adverse physical effects. Broader adoption of IoT devices and AI—in settings such as public utilities and healthcare—will only exacerbate these potential effects.”
Security engineers are working on technologies that can mitigate much of this risk, but many solutions won’t be deployed without government involvement. This is not something that the market can solve. Like data privacy, the risks and solutions are too technical for most people and organizations to understand; companies are motivated to hide the insecurity of their own systems from their customers, their users, and the public; the interconnections can make it impossible to connect data breaches with resultant harms; and the interests of the companies often don’t match the interests of the people.
Governments need to play a larger role: setting standards, policing compliance, and implementing solutions across companies and networks. And while the White House Cybersecurity National Action Plan says some of the right things, it doesn’t nearly go far enough, because so many of us are phobic of any government-led solution to anything.
The next president will probably be forced to deal with a large-scale Internet disaster that kills multiple people. I hope he or she responds with both the recognition of what government can do that industry can’t, and the political will to make it happen.
This essay previously appeared on Vice Motherboard.
EDITED TO ADD (8/11): An essay that agrees with me.
Posted on July 28, 2016 at 5:51 AM •