US Offensive Cyberwar Policy
Today, the United States is conducting offensive cyberwar actions around the world.
More than passively eavesdropping, we’re penetrating and damaging foreign networks for both espionage and to ready them for attack. We’re creating custom-designed Internet weapons, pretargeted and ready to be “fired” against some piece of another country’s electronic infrastructure on a moment’s notice.
This is much worse than what we’re accusing China of doing to us. We’re pursuing policies that are both expensive and destabilizing and aren’t making the Internet any safer. We’re reacting from fear, and causing other countries to counter-react from fear. We’re ignoring resilience in favor of offense.
Welcome to the cyberwar arms race, an arms race that will define the Internet in the 21st century.
Presidential Policy Directive 20, issued last October and released by Edward Snowden, outlines US cyberwar policy. Most of it isn’t very interesting, but there are two paragraphs about “Offensive Cyber Effect Operations,” or OCEO, that are intriguing:
OECO can offer unique and unconventional capabilities to advance US national objectives around the world with little or no warning to the adversary or target and with potential effects ranging from subtle to severely damaging. The development and sustainment of OCEO capabilities, however, may require considerable time and effort if access and tools for a specific target do not already exist.
The United States Government shall identify potential targets of national importance where OCEO can offer a favorable balance of effectiveness and risk as compared with other instruments of national power, establish and maintain OCEO capabilities integrated as appropriate with other US offensive capabilities, and execute those capabilities in a manner consistent with the provisions of this directive.
These two paragraphs, and another paragraph about OCEO, are the only parts of the document classified “top secret.” And that’s because what they’re saying is very dangerous.
Cyberattacks have the potential to be both immediate and devastating. They can disrupt communications systems, disable national infrastructure, or, as in the case of Stuxnet, destroy nuclear reactors; but only if they’ve been created and targeted beforehand. Before launching cyberattacks against another country, we have to go through several steps.
We have to study the details of the computer systems they’re running and determine the vulnerabilities of those systems. If we can’t find exploitable vulnerabilities, we need to create them: leaving “back doors,” in hacker speak. Then we have to build new cyberweapons designed specifically to attack those systems.
Sometimes we have to embed the hostile code in those networks—these are called “logic bombs”—to be unleashed in the future. And we have to keep penetrating those foreign networks, because computer systems always change and we need to ensure that the cyberweapons are still effective.
Like our nuclear arsenal during the Cold War, our cyberweapons arsenal must be pretargeted and ready to launch.
That’s what Obama directed the US Cyber Command to do. We can see glimpses of how effective we are in Snowden’s allegations that the NSA is currently penetrating foreign networks around the world: “We hack network backbones—like huge Internet routers, basically—that give us access to the communications of hundreds of thousands of computers without having to hack every single one.”
The NSA and the US Cyber Command are basically the same thing. They’re both at Fort Meade in Maryland, and they’re both led by Gen. Keith Alexander. The same people who hack network backbones are also building weapons to destroy those backbones. At a March Senate briefing, Alexander boasted of creating more than a dozen offensive cyber units.
Longtime NSA watcher James Bamford reached the same conclusion in his recent profile of Alexander and the US Cyber Command (written before the Snowden revelations). He discussed some of the many cyberweapons the US purchases:
According to Defense News’ C4ISR Journal and Bloomberg Businessweek, Endgame also offers its intelligence clients—agencies like Cyber Command, the NSA, the CIA, and British intelligence—a unique map showing them exactly where their targets are located. Dubbed Bonesaw, the map displays the geolocation and digital address of basically every device connected to the Internet around the world, providing what’s called network situational awareness. The client locates a region on the password-protected web-based map, then picks a country and city—say, Beijing, China. Next the client types in the name of the target organization, such as the Ministry of Public Security’s No. 3 Research Institute, which is responsible for computer security—or simply enters its address, 6 Zhengyi Road. The map will then display what software is running on the computers inside the facility, what types of malware some may contain, and a menu of custom-designed exploits that can be used to secretly gain entry. It can also pinpoint those devices infected with malware, such as the Conficker worm, as well as networks turned into botnets and zombies—the equivalent of a back door left open…
The buying and using of such a subscription by nation-states could be seen as an act of war. ‘If you are engaged in reconnaissance on an adversary’s systems, you are laying the electronic battlefield and preparing to use it’ wrote Mike Jacobs, a former NSA director for information assurance, in a McAfee report on cyberwarfare. ‘In my opinion, these activities constitute acts of war, or at least a prelude to future acts of war.’ The question is, who else is on the secretive company’s client list? Because there is as of yet no oversight or regulation of the cyberweapons trade, companies in the cyber-industrial complex are free to sell to whomever they wish. “It should be illegal,” said the former senior intelligence official involved in cyberwarfare. “I knew about Endgame when I was in intelligence. The intelligence community didn’t like it, but they’re the largest consumer of that business.”
That’s the key question: How much of what the United States is currently doing is an act of war by international definitions? Already we’re accusing China of penetrating our systems in order to map “military capabilities that could be exploited during a crisis.” What PPD-20 and Snowden describe is much worse, and certainly China, and other countries, are doing the same.
All of this mapping of vulnerabilities and keeping them secret for offensive use makes the Internet less secure, and these pretargeted, ready-to-unleash cyberweapons are destabilizing forces on international relationships. Rooting around other countries’ networks, analyzing vulnerabilities, creating back doors, and leaving logic bombs could easily be construed as acts of war. And all it takes is one overachieving national leader for this all to tumble into actual war.
It’s time to stop the madness. Yes, our military needs to invest in cyberwar capabilities, but we also need international rules of cyberwar, more transparency from our own government on what we are and are not doing, international cooperation between governments, and viable cyberweapons treaties. Yes, these are difficult. Yes, it’s a long, slow process. Yes, there won’t be international consensus, certainly not in the beginning. But even with all of those problems, it’s a better path to go down than the one we’re on now.
We can start by taking most of the money we’re investing in offensive cyberwar capabilities and spend them on national cyberspace resilience. MAD, mutually assured destruction, made sense because there were two superpowers opposing each other. On the Internet there are all sorts of different powers, from nation-states to much less organized groups. An arsenal of cyberweapons begs to be used, and, as we learned from Stuxnet, there’s always collateral damage to innocents when they are. We’re much safer with a strong defense than with a counterbalancing offense.
This essay originally appeared on CNN.com. It had the title “Has U.S. Started an Internet War?”—which I had nothing to do with. Almost always, editors choose titles for my essay without asking my opinion—or telling me beforehand.
EDITED TO ADD: Here’s an essay on the NSA’s—or Cyber Command’s—TAO: the Office of Tailored Access Operations. This is the group in charge of hacking China.
According to former NSA officials interviewed for this article, TAO’s mission is simple. It collects intelligence information on foreign targets by surreptitiously hacking into their computers and telecommunications systems, cracking passwords, compromising the computer security systems protecting the targeted computer, stealing the data stored on computer hard drives, and then copying all the messages and data traffic passing within the targeted email and text-messaging systems. The technical term of art used by NSA to describe these operations is computer network exploitation (CNE).
TAO is also responsible for developing the information that would allow the United States to destroy or damage foreign computer and telecommunications systems with a cyberattack if so directed by the president. The organization responsible for conducting such a cyberattack is US Cyber Command (Cybercom), whose headquarters is located at Fort Meade and whose chief is the director of the NSA, Gen. Keith Alexander.
None of this is new. Read this Seymour Hersh article on this subject from 2010.
Subscribe to comments on this entry
h4xx • June 21, 2013 12:04 PM
OpenBSD is the answer. Look how they have completely redesigned giant routers with OpenBGP and pfsense and they just released OpenSMTP finally securing that old garbage mail protocol. Best part is they have kept all configuration uniform so if you know OpenBSD pf firewall syntax you know how to configure OpenSSH, OpenBGP, OpenSMTP ect.
This is the answer to insecure nerworks throw out those corporate blackboxes with NSA back doors and build your own countries infrastructure using OBSD. Run linux work terminals or OBSD terminals that are virtual routed and isolated with carp/pf so they are unable to be attacked or attack other systems on the network. Cost is minimal you just need somebody that knows what they are doing or read the man pages/get the book Absolute Openbsd 2 that just came out.
I don’t understand why its still acceptable to drop cisco or MS blackboxes on a secure network handling stuff like nuclear secrets or critical national telecoms. NSA would have their work cut out for them trying to logic bomb a hardened gentoo or even fedora workstation running grsec and selinux policy, surrounded by a locked down openbsd network firewall and router.