Software [In]security: Cyber War - Hype or Consequences?
According to some of my foreign friends, the United States is wound around the cyber war axle for no apparent reason. They decry the recent spike in coverage of the cyber war issue as exaggerated and irrational. Given the relentless run up to the war in Iraq, perhaps they have a point (and they certainly have every right to be wary of U.S. policy). Yet a careful look at the problem uncovers something that may well be worth worrying about. So who is right?
What is Cyber War Anyway?
One of the main problems with debating and discussing cyber war and its implications is trying to figure out just what cyber war is, anyway. Definitions vary. The "war" part is fairly straightforward (violent conflict between societies for political, economic, or philosophical reasons). But should a definition of cyber war be limited to cyberspace? That is, are we talking about something as simple as taking down somebody's website or infecting their computer with malware (with impact limited to cyberspace)? Probably not. In fact, having tangible impact in the real (non-cyber) world seems to me to be an important part of cyber war. In my view, cyber war requires what war theorists call a "kinetic" aspect to be present. Put another way, an attack in cyberspace needs to have impressive kinetic impact of some sort out in the real world to be considered an act of war. So, if I can infect your command and control system with malware that gives me complete control, and then cause your predator drones to shoot at the wrong targets (a kinetic impact), that would count as an act of cyber war.
Richard A. Clarke, of 9-11 counterterrorism fame, has plenty to say about cyber war. His most recent book Cyber War includes two stories that serve to flesh out my kinetic impact idea further. Clarke presents both stories as acts of cyber war. I'm not sure I agree.
Lets look at the easy case first. The story I found the most riveting and striking in Clarke's book involves Israeli cyber war maneuvers surrounding the bombing of the Syrian nuclear facility in 2007. Syria's formidable air defense system was completely unable to reveal the presence of Israeli F-15s swooping in to bomb the laser-painted target. According to Clarke's book, the air defense system was "p0wned" by Israeli cyber warriors who incapacitated (or otherwise blinded) it before the raid. (Check out the book for details.) According to my definition, this certainly counts as cyber war. The tie to a kinetic result is clear — a completely destroyed Syrian facility.
On the other end of the spectrum (no kinetic result), things are much murkier. Clarke recounts some fairly ho-hum distributed denial of service (dDoS) attacks and labels them acts of cyber war too. In particular, he invokes the Russian dDoS attack against Estonia more than once, and also devotes a few pages to similar events in the Republic of Georgia. Was there enough kinetic impact to count? Not for me. Plus the technical sophistication of those attacks was quite low. It's abundantly clear that similar attacks against popular U.S. e-commerce websites (Amazon or Google) would fizzle to the point of not even being noticed.
I'm not a big fan of mixing cyber script kiddie activity with more serious attacks, because I think it cheapens any budding understanding of the seriousness of cyber warfare and its implications. Because real geeks are always the first people that policy wonks turn to in order to cut through the relentless computer security jargon and hype, it is better not to risk the buy-in of the geeks with a loose definition. DDos with no kinetic impact probably won't cut it. (I know, I know, a dDoS attack on the air traffic control system could cause some problems…but I will appeal to kinetic impact here to salvage my argument.)
When pressed on the mixing of stories issue, Clarke responds that he merely wants to put all of the data on the table and let people decide for themselves. (You can watch me discuss that point with Richard Clarke on Silver Bullet Security Podcast episode 50.)
Of course, war has both defensive and offensive aspects, and understanding this is central to understanding cyber war. An over concentration on offense can be very dangerous. The U.S. is supposedly very good at offense, but from a defense perspective it's a completely different story. As usual, what I call "the NASCAR effect" applies — it is much more entertaining and shiny to talk about offense and its impacts than to focus on defense and building things right. A balanced approach to cyber war describes offense, defense, and exposure in equal measures.
Is Cyber War Inevitable?
Clarke argues that asymmetries and imbalance in offense, defense, and exposure put the United States in an unenviable cyber war position with respect to traditionally weaker potential enemies such as North Korea and Iran. The big problem is that the United States has critical infrastructure that is very much net-centric. Not only does the military practice and preach "net-centric warfare," but the power grid, the air transportation system, the train control system, and telecommunications systems all depend on reasonable cyber security now. Of all these exposures, I believe the power grid is the most worrisome. It's really easy to gin up scary stories.
By contrast with the United States, North Korea barely has a power grid at all, and certainly has very little internet-enabled critical infrastructure. This skews the familiar Cold War concept of "mutually assured destruction" pretty fundamentally.
In my view, it is fairly easy to see how U.S. dependency on internet-connected critical infrastructure is a big problem. The more we count on what is mostly-nonexistent computer security for our non-military infrastructure, the more likely cyber war becomes.
A major challenge for those worried about cyber war is convincing policy-makers, pundits, and the press who covers them to care. You can see this for yourself in the Washington Post's review of Clarke's book. Very real cyber war consequences sound outlandish and unrealistic…like a badly-written Hollywood movie. Those of us who work on computer security all day know that the consequences that Clarke describes are not only realistic — in many cases they may even be understated. The challenge is to convince people that this is an important issue to address without seeming like a lunatic.
We need to determine a sober and clear way to emphasize exposure to cyber war attack. I think some of the stories currently floating around are too sensational and thus seem to exaggerate possible problems even though, when you think through them carefully, they may well explain very real and very important exposures. This puts us in the tricky position of appearing to over-hype what remains undersold.
The Way Forward
You won't be surprised to hear me say that they only way we can lessen both the potential impact and the very possibility of cyber war is to build more secure systems. This starts with the software we all rely on to work. What's the root of smart grid insecurity? Software. Malicious code? Software vulnerability. For what it's worth, Clarke resonates with this view, and makes the point of saying so on page 86 where he says, "Of the three things that make cyber war possible, the most important may be the flaws in software and hardware."
Fortunately we are making important tangible progress on software security. However, I would like to see the palpable sense of urgency that can be observed among major international banks when it comes to computer security spread to the government — and in particular to the civilian side of the government. The Department of Homeland Security has a huge job to do to help defend the nation against cyber attack, and they are doing a particularly namby-pamby job of it. Just for the record, filling out common criteria forms is not at all what I'm talking about.
Clarke also makes a reasoned case near the end of the book for arms control talks focused on cyber war. As one of the principals behind nuclear arms control when he worked for the State Department, one assumes he knows what he's talking about. In any case, it appears that some policy makers are finally beginning to pay attention to cyber war. NATO is even making statements about cyber war and the alliance.
Watch Silver Bullet Episode 50
To celebrate the 50th monthly episode of the Silver Bullet Security Podcast (which I host), we decided to produce a High Definition video of our standard audio interview. Richard Clarke was kind enough to agree to be interviewed. You can find our lengthy discussion of Cyber War (and more about some of the ideas touched on here) on the Silver Bullet Website or on YouTube.