When policy-makers start waving their hands about a definition, you should immediately suspect foul play, and "cyber war" is a term that seems to have almost as many definitions as it has proponents.
An acceptable one is: “Using attacks against an enemy's computer systems to degrade or damage their ability to make war, as part of conflict.” In the current political climate, there are some non-military activities that are sometimes nestled under the umbrella of “cyber war”, including economic espionage and spoiler operations.
In scoping cyber war, I prefer to think of cyber war as a strategic activity. Some like to categorise practically anything involving men in uniform with computers as cyber war; I think, however, that there is adequate terminology such as “tactical battlefield intelligence” to cover the day-to-day use of computers in military operations.
Unfortunately, I think that the real definition mostly likely used is: “If we do it to you, it's cyber war; if you do it to us, it's cyber terrorism.” In other words, cyber war is well on its way to becoming a weapon of privilege – something that we can use on you, but you'd better not even dream of using on us.
It's important to understand this because it reveals a tacit acknowledgement that it's probably not an appropriate weapon to use at all. That's why we consider weapons of privilege like nuclear weapons or armed drones to be somehow outside the normal arsenal of what is fair.
Why? One simple answer, based on Kant's categorical imperative, is that you don't want to use certain weapons on someone else because you are thereby creating the kind of universe in which such weapons are used. In such a universe, you may also be the target of those weapons.
Put differently, if you don't like the idea of being a target of nuclear weapons, you should work toward nuclear weapon bans rather than trying to monopolise all the nuclear weapons for yourself. If you're willing to inflict cyber war, you're helping build a world in which you also become a target of cyber war.
The Geneva Conventions are the closest thing that we have to rules for how to engage in conflict, though they have been consistently ignored in every major conflict since WWI. They still carry weight though – a global acknowledgement that certain actions are never appropriate and those who engage in them are rightly called war criminals.
That brings us to the next question: is cyber war a war crime? One misunderstanding many people have with the Geneva Conventions is that they deal with “declared” wars. They don't, they refer to “conflict”, whether armed or not, whether a declared war, insurrection, civil war or other conflict.
Many of the cyber war proponents I've debated with claim that the Geneva Conventions may not apply because “nobody declares war any more”, which is a shorthand way of saying, “I haven't read the Geneva Conventions and don't know what I'm talking about”. That's why Ratko Mladic is in The Hague right now, for his ordering of massacres of non-combatants during a conflict that was not a declared war.
If you read the Geneva Conventions, you'll realise that mostly they are concerned with protecting non-combatants. We are told over and over that targeting non-combatants is a violation. Collateral damage is not acceptable damage, it is something to be minimised if at all possible, not merely when convenient. So how do we square that with the fact that most cyber-war attacks would involve non-combatant infrastructure, computers and networks? We don't. If a country crashes another country's civilian power grid as part of a cyber attack, it is not acceptable collateral damage.
Cyber war is a strategic proposition and therefore results in the long-term targeting of civilian infrastructure. We acknowledge that we wouldn't find it acceptable if this type of attack was used on us, while the military/intelligence communities coolly talk about such plans – most of us suspect they already have with Stuxnet.
Protocol 2, Article 15 of the Geneva Convention (1977) specifically says: “Works or installations containing dangerous forces, namely dams, dykes and nuclear electrical generating stations, shall not be made the object of attack, even where these objects are military objectives, if such attack may cause the release of dangerous forces and consequent severe losses among the civilian population.”
You can be absolutely certain that if some foreign power were to interfere seriously with a nuclear reactor in the United States, it would be treated as an act of war.
We should be working to establish guidelines for how to de-militarise cyberspace on the basis that cyberspace is almost entirely civilian infrastructure. If the Department of Defence wants to talk about cyber war, they should be acting immediately to stop tunneling their traffic across the civilian data infrastructure so that they can be a proper target.
They then should also encourage their uniformed counterparts around the world to do likewise. We wouldn't want to think that military communications are using civilian networks as a human shield (a violation of the fourth Geneva Convention). If they truly believe that military networks are a legitimate cyber war target, then it's time for them to get off the civilian internet entirely.
I know that my arguments here come off as idealistic, but as someone who has spent his working life trying to defend networks regardless of who they belong to, it pains me greatly to see that cyber space is being claimed as the next battleground for militarists.
Any student of the 20th century knows that the consistent trend of that war-torn century was to increasingly place civilians in the line of fire; that's not a trend any of us should welcome for cyber space either. It's very disappointing to see that one of mankind's brightest achievements for knowledge and globalism is locked in the crosshairs of militarists and want-to-be war criminals.
Marcus Ranum is chief security officer at Tenable Network Security