The new cyber arms race
Tomorrow's wars will be fought not just with guns, but with the click of a mouse half a world away that will unleash weaponized software that could take out everything from the power grid to a chemical plant.
AP Photo/John Kehe staff illustration
Arlington, Va.; and Idaho Falls, Idaho
Deep inside a glass-and-concrete office building in suburban Washington, Sean McGurk grasps the handle of a vault door, clicks in a secret entry code, and swings the steel slab open. Stepping over the raised lip of a submarinelike bulkhead, he enters a room bristling with some of the most sophisticated technology in the United States.
Banks of computers, hard drives humming on desktops, are tied into an electronic filtering system that monitors billions of bits of information flowing into dozens of federal agencies each second. At any given moment, an analyst can pop up information on a wall of five massive television screens that almost makes this feel like Cowboys Stadium in Arlington, Texas, rather than a bland office building in Arlington, Va.
The overriding purpose of all of it: to help prevent what could lead to the next world war.
Specifically, the "Einstein II" system, as it is called, is intended to detect a large cyberattack against the US. The first signs of such an "electronic Pearl Harbor" might include a power failure across a vast portion of the nation's electric grid. It might be the crash of a vital military computer network. It could be a sudden poison gas release at a chemical plant or an explosion at an oil refinery.
Whatever it is, the scores of analysts staffing this new multimillion-dollar "watch and warn" center would, presumably, be able to see it and respond, says Mr. McGurk, the facility director. The National Cybersecurity and Communications Integration Center (NCCIC, pronounced en-kick) is one of the crown jewels of the Department of Homeland Security (DHS). It is linked to four other key watch centers run by the FBI, the Department of Defense (DOD), and the National Security Agency (NSA) that monitor military and overseas computer networks.
They are monuments to what is rapidly becoming a new global arms race. In the future, wars will not just be fought by soldiers with guns or with planes that drop bombs. They will also be fought with the click of a mouse a half a world away that unleashes carefully weaponized computer programs that disrupt or destroy critical industries like utilities, transportation, communications, and energy. Such attacks could also disable military networks that control the movement of troops, the path of jet fighters, the command and control of warships.
"The next time we want to go to war, maybe we wouldn't even need to bomb a country," says Liam O'Murchu, manager of operations for Symantec Security Response, a Mountain View, Calif., computer security firm. "We could just, you know, turn off its power."
In this detached new warfare, soldiers wouldn't be killing other soldiers on the field of battle. But it doesn't mean there might not be casualties. Knocking out the power alone in a large section of the US could sow chaos. What if there were no heat in New England in January? No refrigeration for food? The leak of a radiation plume or chemical gas in an urban area? A sudden malfunction of the stock market? A disrupted air traffic control system?
These are the darkest scenarios, of course – the kind that people spin to sell books and pump up budgets for new cyberwar technology. Interviews with dozens of cyberconflict experts indicate that this kind of strategic, large-scale digital warfare – while possible – is not the most likely to happen. Instead, some see a prolonged period of aggressive cyberespionage, sabotage, and low-level attacks that damage electronic networks. As one recent study done for the Organization for Economic Cooperation and Development put it: "It is unlikely that there will ever be a true cyberwar."
Yet others say that conclusion might be too conservative. The fact is, no one knows for sure where digital weaponry is heading. The cyber arms race is still in its infancy, and once a cybershot is fired, it's hard to predict where the fusillade might end. In the seconds or minutes it might take staffers at the NCCIC to detect an attack, it could have already spread to US water supplies, railway networks, and other vital industries. How does the US military respond – or even know whom to retaliate against? If it does hit back, how does it prevent cyberweapons from spreading damage electronically to other nations around the world?
Policy experts are just beginning to ask some of these questions as the cyberweapons buildup begins. And make no mistake, it is beginning. By one estimate, more than 100 nations are now amassing cybermilitary capabilities. This doesn't just mean erecting electronic defenses. It also means developing "offensive" weapons.
Shrouded in secrecy, the development of these weaponized new software programs is being done outside public view and with little debate about their impact on existing international treaties and on conventional theories of war, like deterrence, that have governed nations for decades.
"Here's the problem – it's 1946 in cyber," says James Mulvenon, a founding member of the Cyber Conflict Studies Association, a nonprofit group in Washington. "So we have these potent new weapons, but we don't have all the conceptual and doctrinal thinking that supports those weapons or any kind of deterrence. Worse, it's not just the US and Soviets that have the weapons – it's millions and millions of people around the world that have these weapons."
In the new cyber world order, the conventional big powers won't be the only ones carrying the cannons. Virtually any nation – or terrorist group or activist organization – with enough money and technical know-how will be able to develop or purchase software programs that could disrupt distant computer networks.
And the US, because it's so wired, is more vulnerable than most big powers to this new form of warfare. It's the price the country may one day pay for being an advanced and open society.
"If the nation went to war today, in a cyberwar, we would lose," Mike McConnell, director of national intelligence from 2007 to 2009, told a US Senate committee a year ago. "We're the most vulnerable. We're the most connected. We have the most to lose."
Still, none of this means people should immediately run for a digital fallout shelter. Many analysts think the cyberwar threat is overblown, and the US is developing sophisticated defenses, such as the digital ramparts here in Arlington. The question is: Will it be enough, or will it all amount to a Maginot line?
The cyber equivalent of the dropping of the atom bomb on Hiroshima came last fall. That's when the world found out about Stuxnet, the software program that wasn't just another annoying virus. It was a sophisticated digital superweapon. Unlike typical malicious software – Trojans and viruses that lurk hidden in a computer to, say, steal a bank account password or some proprietary corporate information – Stuxnet was designed to inflict damage in the real world. In this case it was apparently intended to destroy machines critical to Iran's nuclear ambitions.
The marauding software was introduced into Iranian computers in five locations sometime in 2009, probably, experts believe, by an infected "thumb drive," a portable memory stick, inserted into the network by unwitting Russian engineers who were working on the Iranian nuclear facility. Once inside the system, analysts say, Stuxnet sought out its target, the computer-controlled nuclear centrifuge system, and sabotaged the machinery. Experts believe, in the end, the software may have damaged up to 1,000 of the plant's centrifuges. It did so without any human help – without anyone clicking a mouse or guiding it electronically.
Since its emergence, Stuxnet has demonstrated that cyberattacks will not remain just banal attempts to delete or steal information inside computers or on the Internet. It showed that a cyberweapon can destroy actual plants and equipment – strategically important equipment. It is a "game changer," McGurk told Congress last fall.
Experts believe that Stuxnet was developed by a nation with a top-notch covert cyberweapons team, probably at a cost of millions of dollars. But now that elements of its software code – its electronic blueprint – are available on the Internet, it could be downloaded and reverse-engineered by organized crime groups, cyberweapons dealers, so-called "hactivist" organizations, rogue nations, and terrorists. The hactivist group Anonymous recently touted that it had acquired a copy of the Stuxnet code. Individual tinkerers are getting it, too.
"What Stuxnet represents is a future in which people with the funds will be able to buy a sophisticated attack like this on the black market," says Ralph Langner, a German cyber-security researcher and Stuxnet expert. "Everyone can have their own cyberweapon." He adds that Stuxnet could be modified by someone who isn't even a control-systems expert into a "digital dirty bomb" that could damage or destroy virtually any industrial operating system it targets.
Amr Thabet, an engineering student at the University of Alexandria in Egypt, typifies how easy it is to access the new world of cyberweaponry. During recent mass street protests in his country, he found time to post on his blog a portion of the Stuxnet cyberweapon he had reverse-engineered. The blog drew the attention of cybersecurity experts, who were unhappy, but not surprised, by what he had done.
"This kid's work makes Stuxnet a lot more accessible and portable to other computer architectures," says Bob Radvanovsky, an industrial control-systems expert at Infracritical, a Chicago-based computer security organization. "It's something a number of people are doing for intellectual exercise – or for malicious purposes. It's not a good trend. If a college student is trying to dabble with this, who else on the dark nets with more nefarious intentions might be [as well]?"
In an e-mail interview, Mr. Thabet said he did it largely for the thrill. He noted that he spent two months deconstructing a small but crucial part of the code after he saw all the attention surrounding the discovery of Stuxnet last fall. "It's the first time I see a malware becomes like a gun or like a weapon close a whole company in few days," he writes in broken English. "You can say [Stuxnet] makes the malware a harder challenge and more dangerous. That's maybe what inspire me."
THE 'WAR' HAS ... ALREADY BEGUN?
Definitions of what constitute a "cyberattack" or "cyberwar" vary, but experts roughly agree the US is now immersed in a continuous series of cyberconflicts. These are with state and nonstate actors, from Russia and China to criminal gangs and online protest groups.
"Are we in a cyberwar now?" asks John Bumgarner, research director at the US Cyber Consequences Unit, a Washington-based think tank, who once was a cyberwarrior with the US Army. "No, not yet. Are we being targeted and our nation's networks attacked and infiltrated by nations that may be our adversaries in the future? Yes."
Melissa Hathaway, former acting senior director for cyberspace at the National Security Council, says the threat is less a military one by nation-states and more about the need to protect US intellectual property from spies and organized crime groups.
"We are currently in an economic cyberwar," Ms. Hathaway says. "It is costing our corporations their innovation, costing Americans their jobs, and making us a country economically weaker over the long term. I don't see it emerging as a military conflict, but as an economic war in which malware and our own digital infrastructure is being used to steal our future."
Others agree that a strategic cyberwar isn't likely right now. But they do see the potential for escalation beyond the theft of the latest blueprints for an electric car or jet-fighter engine, particularly as the technology of digital warfare advances and becomes a more strategic imperative.
"We in the US tend to think of war and peace as an on-off toggle switch – either at full-scale war or enjoying peace," says Joel Brenner, former head of counterintelligence under the US Director of National Intelligence. "The reality is different. We are now in a constant state of conflict among nations that rarely gets to open warfare.... What we have to get used to is that even countries like China, with which we are certainly not at war, are in intensive cyberconflict with us."
While he agrees the notion of big-scale cyberwarfare has been over-hyped, he says attacks that move beyond aggressive espionage to strikes at, or sabotage of, industrial processes and military systems "will become a routine reality."
ANYTHING YOU CAN DO, WE CAN DO BETTER
The attacks were coordinated but relatively unsophisticated: In the spring of 2007, hackers blocked the websites of the Estonian government and clogged the country's Internet network. At one point, bank cards were immobilized. Later, in 2008, similar cyberstrikes preceded the Russian invasion of Georgia. Moscow denied any involvement in the attacks, but Estonia, among others, suspected Russia.
Whoever it was may not be as important as what it's done: touched off a mini cyber arms race, accelerated by the Stuxnet revelation.
Germany and Britain announced new cybermilitary programs in January. In December, Estonia and Iran unveiled cybermilitias to help defend against digital attack. They join at least 20 nations that now have advanced cyberwar programs, according to McAfee, a Santa Clara, Calif., computer security firm. Yet more than 100 countries have at least some cyberconflict prowess, and multiple nations "have the capability to conduct sustained, high-end cyberattacks against the US," according to a new report by the Cyber Conflict Studies Association.
McAfee identifies a handful of countries moving from a defensive to a more offensive posture – including the US, China, Russia, France, and Israel. Experts like Mr. Langner say the US is the world's cyber superpower, with weapons believed to be able to debilitate or destroy targeted computer networks and industrial plants and equipment linked to them. Indeed, China widely assumes that their nation's computer systems have been "thoroughly compromised" by the US, according to Dr. Mulvenon of the Cyber Conflict Studies Association, even as the Chinese penetrate deeper into US industrial and military networks.
As well armed as the US is, however, its defenses are porous. The US may have the mightiest military in the world, but it is also the most computerized – everything from smart bombs to avionics to warship controls – making it unusually vulnerable to cyberassault.
The DOD's communication system includes some 15,000 computer networks and 7 million computing devices. According to the Pentagon, unknown attackers try to breach its systems 6 million times a day. More than a few attempts have succeeded.
Hackers are believed to have stolen key elements of the F-35 jet fighter a few years ago from a defense contractor. In 2008, infiltrators used thumb drives to infect the DOD's classified electronic network, resulting in what Deputy Defense Secretary William Lynn later called the "most significant breach of US military computers ever."
Unlike many of its potential adversaries, the Pentagon is heavily reliant on computer networks. Over the past two decades, US industry, along with the military and federal agencies, have linked some networks and elements of the nation's infrastructure – power plants, air traffic control systems, rail lines – to the notoriously insecure Internet. It makes it easier, faster, and cheaper to communicate and conduct business – but at a cost. Almost all electrical power used by US military bases, for instance, comes from commercial utilities, and the power grid is a key target of adversaries.
"We're pretty vulnerable today," says a former US national security official. "Our defense is superporous against anything sophisticated."
Countries that are less wired are less vulnerable, which represents another danger. Some analysts even suggest that a small power like North Korea could do serious damage to the US in a cyberattack while sustaining relatively little itself. In a report presented at a NATO conference, former NSA expert Charlie Miller estimated that Pyongyang would need only about 600 cyber experts, three years, and $50 million to overtake and defeat America in a digital war.
"One of North Korea's biggest advantages is that it has hardly any Internet-connected infrastructure to target," he says. "On the other hand, the US has tons of vulnerabilities a country like North Korea could exploit."
I SPY, THE SEQUEL
The elite group of hackers sit at an oval bank of computers in a second-floor office on the wind-swept plains of Idaho. Their mission: infiltrate the computer network of Acme Products, an American industrial plant. They immediately begin probing for ways around the company's cyberdefenses and fire walls. Within minutes, they tap into the plant's electronic controls, sabotaging the manufacturing process.
"They're already inside our system," howls an Acme worker, looking at his unresponsive computer after only 20 minutes. "They've got control of the lights. We can't even control our own lights!"
Less than a half-hour later, a plastic vat is overflowing, spraying liquid into an industrial sink. The company's attempts to retake control of the system prove futile. Is the leak a toxic chemical? Something radioactive?
Fortunately, in this case it is water, and the company itself is fictitious. This is simply an exercise by members of the DHS's Industrial Control System-Computer Emergency Readiness Team (ICS-CERT), simulating an attack and defense of a company.
The message to emerge from the war game is unmistakably clear: Industrial America isn't well prepared for the new era of cyberwar, either.
"We conduct these training classes to alert industry to what's really going on and educate them as to vulnerabilities they may not have thought of," says a senior manager at the Idaho National Laboratory (INL) in Idaho Falls, where the readiness team is located.
Down the street, in another warehouselike building, high walls and locked doors shroud rooms where commercial vendors bring their industrial-control software to be probed for weaknesses by the cyber teams.
Despite all the efforts here, experts say gaping holes exist in America's commercial electronic defenses. One reason is the vast number of people and organizations trying to penetrate the networks of key industries. Some people liken the intensity of the spying to the height of the postwar rivalry between the US and the Soviet Union – only the snooping now isn't just by a few countries.
"I personally believe we're in the middle of a kind of cyber cold war," says a senior industrial control systems security expert at INL. "Over the past year our team has visited 30 to 40 companies in critical infrastructure industries – looking for threats on their [networks and industrial-control] systems – to see the level of penetration. In every case, teams of professionals were already there, embedded on every system."
If only part of this infiltration turned out to be corporate espionage, that would be bad enough. But there's a more insidious threat lurking underneath. In his book "Cyber War," Richard Clarke, former counterterrorism chief with the National Security Council, writes that foreign nations are "preparing the battlefield" in key US industries and military networks, in part by creating "trapdoors" in electronic industrial-control systems.
These trapdoors, in the form of nearly invisible software "rootkits," are designed to give the attacker access and control over industries' computer networks, which could later be used to disrupt or destroy operations – for instance, of the US power grid.
"These hackers are invading the grid's control systems right now where it's easiest, getting themselves in position where they could control things if they wanted to," says the senior cybersecurity expert. "But they're not controlling them yet."
Michael Assante, a former Navy cyberwarfare specialist and INL industrial-security expert, sees calculated hacking taking place as well. "I agree we have a lot of cyberespionage going on and a lot of preparation of the battlefield," he says in an interview at his home on a butte overlooking Idaho's Snake River Valley. "There's no question the grid is vulnerable."
THE GENIE IS OUT OF THE HARD DRIVE
Despite their dangers, cyberweapons hold clear appeal to the US and other nations. For one thing, they don't involve shooting people or inflicting casualties in a conventional sense. If fewer people die from bombs and bullets as a result of surreptitious software programs, nations may be more inclined to use them to try to deal with intractable problems. Cyberweapons may also be far cheaper than many conventional weapons.
No doubt these are among the reasons President Obama has accelerated the development of US cybersecurity efforts, building on programs begun late in the tenure of President George W. Bush. In 2009, when announcing the new position of cybersecurity coordinator, Mr. Obama called digital infrastructure a "strategic national asset." Then, last spring, the Pentagon unveiled its joint US Cyber Command to accelerate and consolidate its digital warfare capabilities – including the ability to strike preemptively. Cyberspace was added to sea, air, land, and space as the fifth domain in which the US seeks "dominance."
"Given the dominance of offense in cyberspace, US defenses need to be dynamic," wrote Mr. Lynn in Foreign Affairs magazine. "Milliseconds can make a difference, so the US military must respond to attacks as they happen or even before they arrive."
Yet the digital war buildup could have far-reaching – and unexpected – consequences. Cyberweapons are hardly clinical or benign. They can infect systems globally in minutes that were not the intended target. Experts say Stuxnet, a self-propagating "worm," corrupted more than 100,000 Windows-based computers worldwide. Its damage could have been far more widespread if the digital warhead had been written to activate on any industrial-control system it found instead of just the one it targeted in Iran.
Because strikes and counterstrikes can happen in seconds, conflicts could quickly escalate outside the world of computers. What, for instance, would the US do if an adversary knocked out a power plant – would it retaliate with digital soldiers or real ones? NATO and other organizations are already weighing whether to respond militarily against nations that launch or host cyberattacks against member states.
"The US cybersecurity strategy since 2003 has stated that we're not just going to respond to cyberattacks with cyber," says Greg Rattray, a former director of cybersecurity for the National Security Council. "If somebody cripples the US electric grid, a nuclear power plant, or starts to kill people with cyberattacks, we have reserved the right to retaliate by the means we deem appropriate."
Yet figuring out whom to retaliate against is far more complicated in a cyberwar than a conventional war. It's not just a matter of seeing who dropped the bombs. The Internet and the foggy world of cyberspace provide ample opportunity for anonymity.
The US and other countries are working on technical systems that would allow them to reverse-engineer attacks, detecting identifying elements among tiny packets of information that bounce among servers worldwide. Yet even if cybersleuths can trace the source of a strike to an individual computer, it might be located in the US. Foreign governments could send elite hackers into other countries to infiltrate networks, making it harder to follow the electronic trail.
"Access is the key thing," says Dr. Brenner, the former counterintelligence chief. "If we ever get to real hostilities, all these attacks are going to be launched from within the US...."
All this makes it difficult to apply conventional doctrines of war, such as deterrence and first-strike capability, to the new era of cyberconflict. Does the US retaliate if it's unsure of who the enemy is? Can there be deterrence if retaliation is uncertain? There are more mundane questions, too: When does aggressive espionage cross a threshold and constitute an "attack"?
"We live in a glass house so we better be careful about throwing rocks," says Hathaway of America's presumed prowess in offensive cyberwar and espionage tactics. "We don't have the resilience built into our infrastructure today to enter into such an escalated environment."
In the face of such ambiguity, many experts say the US needs an overarching policy that governs the use of cyberweapons.
On the plus side, multiple cyberattack technologies "greatly expand the range of options available to US policy makers as well as the policy makers of other nations...," the National Academy of Sciences concluded in a landmark 2009 study. On the other hand, "today's policy and legal framework for guiding and regulating the US use of cyberattack is ill-formed, undeveloped, and highly uncertain."
THE e-MAGINOT LINE
The NCCIC staffers toiling away in their war room in Arlington do face a daunting task. The powerful Einstein II system sifts millions of attacks raining down on federal computer networks each day. The unit sends out alerts and may intervene to stop a penetration.
But can it and other federal "watch centers" really protect the country from a major cyberassault? Perhaps – if they are actually watching.
The fact is, some 85 percent of the computer networks of critical US industries – water systems, stock markets, the US power grid – lie in the hands of private industry and are not monitored directly by federal agencies, McGurk acknowledges. NCCIC's mission is to safeguard government first, then private industry, if it can figure out what's happening. For the power grid, for instance, NCCIC relies on the North American Electric Reliability Corporation and individual utilities to relay information about security breaches.
Even the US military, which operates at least two large watch centers, has "no situational awareness [in cyberspace] – it's very limited," Gen. Keith Alexander, who heads the Pentagon's new US Cyber Command, admitted at a conference last June.
Still, for all the nation's vulnerabilities, people may not want to panic about the digital arms race just yet.
"Some of the greatest minds of our times were able to bottle up nuclear weapons and keep them in their silos with diplomacy," says Mr. Assante. "I think, I hope, something like that will happen with cyber. We can learn to manage the risk ... not bury our heads in the sand."