Entries Tagged "attribution"

Page 3 of 3

Attack Attribution in Cyberspace

When you’re attacked by a missile, you can follow its trajectory back to where it was launched from. When you’re attacked in cyberspace, figuring out who did it is much harder. The reality of international aggression in cyberspace will change how we approach defense.

Many of us in the computer-security field are skeptical of the US government’s claim that it has positively identified North Korea as the perpetrator of the massive Sony hack in November 2014. The FBI’s evidence is circumstantial and not very convincing. The attackers never mentioned the movie that became the centerpiece of the hack until the press did. More likely, the culprits are random hackers who have loved to hate Sony for over a decade, or possibly a disgruntled insider.

On the other hand, most people believe that the FBI would not sound so sure unless it was convinced. And President Obama would not have imposed sanctions against North Korea if he weren’t convinced. This implies that there’s classified evidence as well. A couple of weeks ago, I wrote for the Atlantic, “The NSA has been trying to eavesdrop on North Korea’s government communications since the Korean War, and it’s reasonable to assume that its analysts are in pretty deep. The agency might have intelligence on the planning process for the hack. It might, say, have phone calls discussing the project, weekly PowerPoint status reports, or even Kim Jong Un’s sign-off on the plan. On the other hand, maybe not. I could have written the same thing about Iraq’s weapons-of-mass-destruction program in the run-up to the 2003 invasion of that country, and we all know how wrong the government was about that.”

The NSA is extremely reluctant to reveal its intelligence capabilities—or what it refers to as “sources and methods”—against North Korea simply to convince all of us of its conclusion, because by revealing them, it tips North Korea off to its insecurities. At the same time, we rightly have reason to be skeptical of the government’s unequivocal attribution of the attack without seeing the evidence. Iraq’s mythical weapons of mass destruction is only the most recent example of a major intelligence failure. American history is littered with examples of claimed secret intelligence pointing us toward aggression against other countries, only for us to learn later that the evidence was wrong.

Cyberspace exacerbates this in two ways. First, it is very difficult to attribute attacks in cyberspace. Packets don’t come with return addresses, and you can never be sure that what you think is the originating computer hasn’t itself been hacked. Even worse, it’s hard to tell the difference between attacks carried out by a couple of lone hackers and ones where a nation-state military is responsible. When we do know who did it, it’s usually because a lone hacker admitted it or because there was a months-long forensic investigation.

Second, in cyberspace, it is much easier to attack than to defend. The primary defense we have against military attacks in cyberspace is counterattack and the threat of counterattack that leads to deterrence.

What this all means is that it’s in the US’s best interest to claim omniscient powers of attribution. More than anything else, those in charge want to signal to other countries that they cannot get away with attacking the US: If they try something, we will know. And we will retaliate, swiftly and effectively. This is also why the US has been cagey about whether it caused North Korea’s Internet outage in late December.

It can be an effective bluff, but only if you get away with it. Otherwise, you lose credibility. The FBI is already starting to equivocate, saying others might have been involved in the attack, possibly hired by North Korea. If the real attackers surface and can demonstrate that they acted independently, it will be obvious that the FBI and NSA were overconfident in their attribution. Already, the FBI has lost significant credibility.

The only way out of this, with respect to the Sony hack and any other incident of cyber-aggression in which we’re expected to support retaliatory action, is for the government to be much more forthcoming about its evidence. The secrecy of the NSA’s sources and methods is going to have to take a backseat to the public’s right to know. And in cyberspace, we’re going to have to accept the uncomfortable fact that there’s a lot we don’t know.

This essay previously appeared in Time.

Posted on January 8, 2015 at 6:34 AM • View Comments

Attributing the Sony Attack

No one has admitted taking down North Korea’s Internet. It could have been an act of retaliation by the US government, but it could just as well have been an ordinary DDoS attack. The follow-on attack against Sony PlayStation definitely seems to be the work of hackers unaffiliated with a government.

Not knowing who did what isn’t new. It’s called the “attribution problem,” and it plagues Internet security. But as governments increasingly get involved in cyberspace attacks, it has policy implications as well. Last year, I wrote:

Ordinarily, you could determine who the attacker was by the weaponry. When you saw a tank driving down your street, you knew the military was involved because only the military could afford tanks. Cyberspace is different. In cyberspace, technology is broadly spreading its capability, and everyone is using the same weaponry: hackers, criminals, politically motivated hacktivists, national spies, militaries, even the potential cyberterrorist. They are all exploiting the same vulnerabilities, using the same sort of hacking tools, engaging in the same attack tactics, and leaving the same traces behind. They all eavesdrop or steal data. They all engage in denial-of-service attacks. They all probe cyberdefences and do their best to cover their tracks.

Despite this, knowing the attacker is vitally important. As members of society, we have several different types of organizations that can defend us from an attack. We can call the police or the military. We can call on our national anti-terrorist agency and our corporate lawyers. Or we can defend ourselves with a variety of commercial products and services. Depending on the situation, all of these are reasonable choices.

The legal regime in which any defense operates depends on two things: who is attacking you and why. Unfortunately, when you are being attacked in cyberspace, the two things you often do not know are who is attacking you and why. It is not that everything can be defined as cyberwar; it is that we are increasingly seeing warlike tactics used in broader cyberconflicts. This makes defence and national cyberdefence policy difficult.

In 2007, the Israeli Air Force bombed and destroyed the al-Kibar nuclear facility in Syria. The Syrian government immediately knew who did it, because airplanes are hard to disguise. In 2010, the US and Israel jointly damaged Iran’s Natanz nuclear facility. But this time they used a cyberweapon, Stuxnet, and no one knew who did it until details were leaked years later. China routinely denies its cyberespionage activities. And a 2009 cyberattack against the United States and South Korea was blamed on North Korea even though it may have originated from either London or Miami.

When it’s possible to identify the origins of cyberattacks­—like forensic experts were able to do with many of the Chinese attacks against US networks­—it’s as a result of months of detailed analysis and investigation. That kind of time frame doesn’t help at the moment of attack, when you have to decide within milliseconds how your network is going to react and within days how your country is going to react. This, in part, explains the relative disarray within the Obama administration over what to do about North Korea. Officials in the US government and international institutions simply don’t have the legal or even the conceptual framework to deal with these types of scenarios.

The blurring of lines between individual actors and national governments has been happening more and more in cyberspace. What has been called the first cyberwar, Russia vs. Estonia in 2007, was partly the work of a 20-year-old ethnic Russian living in Tallinn, and partly the work of a pro-Kremlin youth group associated with the Russian government. Many of the Chinese hackers targeting Western networks seem to be unaffiliated with the Chinese government. And in 2011, the hacker group Anonymous threatened NATO.

It’s a strange future we live in when we can’t tell the difference between random hackers and major governments, or when those same random hackers can credibly threaten international military organizations.

This is why people around the world should care about the Sony hack. In this future, we’re going to see an even greater blurring of traditional lines between police, military, and private actions as technology broadly distributes attack capabilities across a variety of actors. This attribution difficulty is here to stay, at least for the foreseeable future.

If North Korea is responsible for the cyberattack, how is the situation different than a North Korean agent breaking into Sony’s office, photocopying a lot of papers, and making them available to the public? Is Chinese corporate espionage a problem for governments to solve, or should we let corporations defend themselves? Should the National Security Agency defend US corporate networks, or only US military networks? How much should we allow organizations like the NSA to insist that we trust them without proof when they claim to have classified evidence that they don’t want to disclose? How should we react to one government imposing sanctions on another based on this secret evidence? More importantly, when we don’t know who is launching an attack or why, who is in charge of the response and under what legal system should those in charge operate?

We need to figure all of this out. We need national guidelines to determine when the military should get involved and when it’s a police matter, as well as what sorts of proportional responses are available in each instance. We need international agreements defining what counts as cyberwar and what does not. And, most of all right now, we need to tone down all the cyberwar rhetoric. Breaking into the offices of a company and photocopying their paperwork is not an act of war, no matter who did it. Neither is doing the same thing over the Internet. Let’s save the big words for when it matters.

This essay previously appeared on TheAtlantic.com.

Jack Goldsmith responded to this essay.

Posted on January 7, 2015 at 11:16 AM • View Comments

1 2 3

Sidebar photo of Bruce Schneier by Joe MacInnis.