Russia Wanted to be Caught, Says Company Waging War on the DNC Hackers

The Russian groups behind the DNC hack no longer seem to care about getting caught. Long before the Kremlin-sponsored hacking squads APT 28 and APT 29 were making waves for stealing files from the Democratic National Committee, they made an appearance in two white papers put out by FireEye. The cybersecurity company has been monitoring and analyzing the two groups on behalf of corporate clients for years. In the DNC breach, a company spokesman told Defense One: “They wanted experts and policymakers to know that Russia is behind it.”

That fits a pattern of increasing bold moves over the past year by the groups, which are also known as FANCY BEAR and COZY BEAR, says Christopher Porter, the manager of Horizons, the strategic intelligence and forecasting arm of FireEye iSIGHT Intelligence, the company’s threat monitoring division.

“We see them now persisting even when they know that security professionals have been called in to remove them from a system. They continue their operational pace at a very high level. So that’s a huge risk and a sea change in their behavior,” Porter said. “Even when they know they’re caught, they don’t stop the operation, necessarily.”

That’s highly unusual for an advanced persistent threat group. It signals that Russia is willing to work in a space normally reserved for criminals, devoting government resources and acting with impunity. That makes them incredibly difficult to counter, for the same reason the West had no good response to the “little green men” — the Russian forces that invaded Ukraine disguised as a organic populist militant movement.

That camouflaged brazenness was also seen in the 2015 hack on the Joint Chiefs of Staff’s nonclassified email, also attributed to APT 29. The attackers were “jumping from one computer to another” in the network, according to a representative of the company the Pentagon hired to fix the damage. “A lot of the time you don’t have the command-and-control architecture to be able to go in and see the attack,” he said. “So the advanced threat characteristics change to be more automated, a kind of pervasive deployment using common vulnerabilities and exploiting them widely.”

That means that APT 29 has stopped retreating from networks when they think they’ve been detected. Now they adapt the hack in the open, bobbing and weaving like a fast and clever boxer, taunting the victim to expel them.

“We’ve even seen them on some systems where they know that there is anti-virus [software] on a computer inside of a network system that they’re on,” FireEye’s Porter said. “They’re moving laterally within a network. They know that their tool is going to be detected by a system that they’re about to move to and they'll do it anyway because they’re such skilled hackers that they can compromise the system and then jump to another system and get what they need before they can be quarantined.”  

There’s a reason that’s not normal behavior, even among very skilled hackers. After attackers are expelled from a system, defenders move quickly to patch the security hole they used. Groups that run advanced persistent attacks move stealthily, lest they burn too quickly through their bag of tricks.

Yet FireEye found that APT 28 and APT 29 didn’t even bother to change the pace of their attacks as their targets became aware of them.

“We have a Mandiant arm that can go back and recreate what happened,” after a breach, Porter said. “When we look back on it over time, there’s no evidence that if their operations were exposed on Tuesday that, on Wednesday, exploitation pace against their targets would change. It didn’t make any difference. They have an armory of zero days,” attacks that have never been seen before.

Case in point: a July 2015 incident in which a security firm published a blog post about how APT 28 was using a specific zero-day exploit. The group updated the hack the next day, as FireEye focused reporting team manager Kristen Dennesen told the RSA conference this year.

Porter thinks that’s one piece of evidence that both groups have state sponsorship. You need more than than coding chops to pull off a stunt like that; it helps to have an international intelligence collection network you can work with.

“If these state-backed actors have professional military or intelligence operators overseeing the operation, any change you can make, they’re going to try and find a counter to that,” he said. “They seem to know that certain white papers are going to be public and they make the changes the day before they come out. We’ve seen evidence that they’ve known in advance that someone is going to reveal that they were going to be discovered and they make changes so that they continue uninterrupted.”

Over the past week, U.S. intelligence community officials have said that they have “high confidence” that the Russian government was behind the theft of emails from the DNC. That’s an unusually bold statement for the IC to make about a data breach that’s currently moving the news cycle. By contrast, the intelligence community still hasn’t made a formal declaration of attribution about the OPM hack. Months after the intrusion was revealed, Clapper acknowledged only that China was the “leading suspect.”

Porter believes that part of the reason that the IC and multiple cyber security researchers were able to implicate Russia is that Russia was showing off. Consider that on June 15, one day after Crowdstrike fingered APT 28 and APT 29, a figure named Guccifer 2.0 claimed to have done the hack, alone. But Twitter users quickly found metadata in Guccifer 2.0’s files that undermined that claim. The docs contained a tag reading “Феликс Эдмундович,” a reference to to the founder of the Soviet Secret Police.

But security expert Jeff Carr thought the smoke off this smoking-gun was a bit too thick. In his minority report, he asks: what kind of spy ring tags their stolen docs before releasing them under a cover?

“Raise your hand if you think that a GRU or FSB officer would add Iron Felix’s name to the metadata of a stolen document before he released it to the world while pretending to be a Romanian hacker. Someone clearly had a wicked sense of humor,” he wrote.

You’ll Never Get 100% Certainty in Cyberwar

This shows the effectiveness of information warfare at this moment in history: a malware attack is fundamentally different from a missile strike that can be seen from space and immediately attributed to a party, a unit, a fixed position on earth, and a piece of machinery (if not an individual operator). One hundred percent certainty in any information attack will always be next to impossible, and that makes it hard to shape policy, legislation, or retribution.

“For U.S. policy makers and a lot of private-sector companies, they tend to be dismissive. They say, ‘Oh, we had a thousand spearphish attacks today.’ The fact that there’s such a huge background noise level in fairly sophisticated cyber crime across so many targets around the world, it allows APT groups to blend in if they want to,” Porter said. “It’s death by a thousand cuts from the perspective of U.S. policy. Any individual cyber criminal act is not a national security concern, but taken in the aggregate, having a high level of cyber crime in general should be a very high-level concern.”

That plays into Russia’s hands. State actors can use headlines about persistent criminal cyber threats to make geopolitical activity look merely criminal.

“If you were to reduce the very high level of cybercrime, states wouldn’t be able to carry out these attacks. They would lose this plausible deniability and it would become a more straightforward attack. I think they want to make it difficult for leaders to have the kind of unambiguous statement that drives policy in a democracy,” Porter said. “They want to make it hard to respond. But they probably don’t mind getting caught, in the sense that they want to send a message.”

Crowdstrike president Shawn Henry is dubious. “I don’t know what kind of foreign intelligence service conducting a covert operation wants to be found,” he said on Thursday, but added that CrowdStrike picked up the DNC hack within 48 hours and that it “wasn’t difficult.”

If you buy Porter’s theory, the question becomes: what kind of message could the Russians mean to send? The FireEye employee guesses that these sorts of breaches are likely a demonstration of capability, or perhaps a reprisal against the West for sanctions against Russian leaders. It’s an idea that he’s sharing with his private-sector customers.

“I view their activities as, they want to muddy the political response in democracies by making it seem like a complicated and ambiguous issue,” Porter said. If they’re willing to do A, B, and C then you need to understand that it’s not difficult for them to target an individual. That’s what cyber gives them. From Russia, they can pick an individual that they want to bully, using the full resources of a state organization. And that’s unprecedented. So if they decide that they want to pick on a certain corporate executive, maybe they could do a particular, hacktivist style leak. Activists go after companies all the time…it’s hard for a the company to prove that their loss was caused by a state and not by a criminal. So the policy is still complicated. That’s a nice place to be in if you’re Russia.”