Where Schneier had sought one overarching technical fix, hard experience had taught him the quest was illusory. Indeed, yielding to the American penchant for all-in-one high-tech solutions can make us less safe—especially when it leads to enormous databases full of confidential information. Secrecy is important, of course, but it is also a trap. The more secrets necessary to a security system, the more vulnerable it becomes.
To forestall attacks, security systems need to be small-scale, redundant, and compartmentalized. Rather than large, sweeping programs, they should be carefully crafted mosaics, each piece aimed at a specific weakness. The federal government and the airlines are spending millions of dollars, Schneier points out, on systems that screen every passenger to keep knives and weapons out of planes. But what matters most is keeping dangerous passengers out of airline cockpits, which can be accomplished by reinforcing the door. Similarly, it is seldom necessary to gather large amounts of additional information, because in modern societies people leave wide audit trails. The problem is sifting through the already existing mountain of data. Calls for heavy monitoring and record-keeping are thus usually a mistake. ("Broad surveillance is a mark of bad security," Schneier wrote in a recent Crypto-Gram.)
To halt attacks once they start, security measures must avoid being subject to single points of failure. Computer networks are particularly vulnerable: once hackers bypass the firewall, the whole system is often open for exploitation. Because every security measure in every system can be broken or gotten around, failure must be incorporated into the design. No single failure should compromise the normal functioning of the entire system or, worse, add to the gravity of the initial breach. Finally, and most important, decisions need to be made by people at close range—and the responsibility needs to be given explicitly to people, not computers.
Unfortunately, there is little evidence that these principles are playing any role in the debate in the Administration, Congress, and the media about how to protect the nation. Indeed, in the argument over policy and principle almost no one seems to be paying attention to the practicalities of security—a lapse that Schneier, like other security professionals, finds as incomprehensible as it is dangerous.
A couple of months after September 11, I flew from Seattle to Los Angeles to meet Schneier. As I was checking in at Sea-Tac Airport, someone ran through the metal detector and disappeared onto the little subway that runs among the terminals. Although the authorities quickly identified the miscreant, a concession stand worker, they still had to empty all the terminals and re-screen everyone in the airport, including passengers who had already boarded planes. Masses of unhappy passengers stretched back hundreds of feet from the checkpoints. Planes by the dozen sat waiting at the gates. I called Schneier on a cell phone to report my delay. I had to shout over the noise of all the other people on their cell phones making similar calls. "What a mess," Schneier said. "The problem with airport security, you know, is that it fails badly."
For a moment I couldn't make sense of this gnomic utterance. Then I realized he meant that when something goes wrong with security, the system should recover well. In Seattle a single slip-up shut down the entire airport, which delayed flights across the nation. Sea-Tac, Schneier told me on the phone, had no adequate way to contain the damage from a breakdown—such as a button installed near the x-ray machines to stop the subway, so that idiots who bolt from checkpoints cannot disappear into another terminal. The shutdown would inconvenience subway riders, but not as much as being forced to go through security again after a wait of several hours. An even better idea would be to place the x-ray machines at the departure gates, as some are in Europe, in order to scan each group of passengers closely and minimize inconvenience to the whole airport if a risk is detected—or if a machine or a guard fails.
Schneier was in Los Angeles for two reasons. He was to speak to ICANN, the Internet Corporation for Assigned Names and Numbers, which controls the "domain name system" of Internet addresses. It is Schneier's belief that attacks on the address database are the best means of taking down the Internet. He also wanted to review Ginza Sushi-Ko, perhaps the nation's most exclusive restaurant, for the food column he writes with his wife, Karen Cooper.
Minutes after my delayed arrival Schneier had with characteristic celerity packed himself and me into a taxi. The restaurant was in a shopping mall in Beverly Hills that was disguised to look like a collection of nineteenth-century Italian villas. By the time Schneier strode into the tiny lobby, he had picked up the thread of our airport discussion. Failing badly, he told me, was something he had been forced to spend time thinking about.
In his technophilic exuberance he had been seduced by the promise of public-key encryption. But ultimately Schneier observed that even strong crypto fails badly. When something bypasses it, as the keystroke logger did with Nicodemo Scarfo's encryption, it provides no protection at all. The moral, Schneier came to believe, is that security measures are characterized less by their manner of success than by their manner of failure. All security systems eventually miscarry. But when this happens to the good ones, they stretch and sag before breaking, each component failure leaving the whole as unaffected as possible. Engineers call such failure-tolerant systems "ductile." One way to capture much of what Schneier told me is to say that he believes that when possible, security schemes should be designed to maximize ductility, whereas they often maximize strength.
Since September 11 the government has been calling for a new security infrastructure—one that employs advanced technology to protect the citizenry and track down malefactors. Already the USA PATRIOT Act, which Congress passed in October, mandates the establishment of a "cross-agency, cross-platform electronic system ... to confirm the identity" of visa applicants, along with a "highly secure network" for financial-crime data and "secure information sharing systems" to link other, previously separate databases. Pending legislation demands that the Attorney General employ "technology including, but not limited to, electronic fingerprinting, face recognition, and retinal scan technology." The proposed Department of Homeland Security is intended to oversee a "national research and development enterprise for homeland security comparable in emphasis and scope to that which has supported the national security community for more than fifty years"—a domestic version of the high-tech R&D juggernaut that produced stealth bombers, smart weapons, and anti-missile defense.
Iris, retina, and fingerprint scanners; hand-geometry assayers; remote video-network surveillance; face-recognition software; smart cards with custom identification chips; decompressive baggage checkers that vacuum-extract minute chemical samples from inside suitcases; tiny radio implants beneath the skin that continually broadcast people's identification codes; pulsed fast-neutron analysis of shipping containers ("so precise," according to one manufacturer, "it can determine within inches the location of the concealed target"); a vast national network of interconnected databases—the list goes on and on. In the first five months after the terrorist attacks the Pentagon liaison office that works with technology companies received more than 12,000 proposals for high-tech security measures. Credit-card companies expertly manage credit risks with advanced information-sorting algorithms, Larry Ellison, the head of Oracle, the world's biggest database firm, told The New York Times in April; "We should be managing security risks in exactly the same way." To "win the war on terrorism," a former deputy undersecretary of commerce, David J. Rothkopf, explained in the May/June issue of Foreign Policy, the nation will need "regiments of geeks"—"pocket-protector brigades" who "will provide the software, systems, and analytical resources" to "close the gaps Mohammed Atta and his associates revealed."