May 15, 2001

by Bruce Schneier
Founder and CTO
Counterpane Internet Security, Inc.

A free monthly newsletter providing summaries, analyses, insights, and commentaries on computer security and cryptography.

Back issues are available at <>. To subscribe or unsubscribe, see below.

Copyright (c) 2001 by Counterpane Internet Security, Inc.

In this issue:

Defense Options: What Military History Can Teach Network Security, Part 2

In Part I of this series, I examined the natural advantages of defense in military history. I concluded that two advantages—the ability to shift forces and knowledge of the terrain—are underutilized in network security. I concluded that network security based on hidden attack sensors and rapid response would be far more effective than firewalls, IDSs, and whatever the new new thing next new thing is.

In Part II, I want to look even more broadly at the military’s notion of defense. In war, there are three, and only three, types of defense: passive defense, active defense, and counterattack.

Passive defenses involve making yourself harder to attack. Against an air assault, for example, this could mean building bunkers or hiding in caves, dispersing your forces, or covering yourself in camouflage. All of these defenses have the same goal: reducing the effectiveness of the enemy’s bombs. The important thing to note is that while passive defenses make attacks less effective, they do nothing to the attackers themselves.

Active defenses are designed to take out the attacker. Returning to the incoming aircraft example, an active attack could be anti-aircraft fire that shoots down the attacking aircraft in flight. This is harder than passive defense, but can be much more effective.

Counterattack means turning the tables and attacking the attacker. Against the air assault, it could involve attacking airfields, fuel depots, and ammunition storage facilities. Note that the line between defense and offense can blur, as some counterattack targets are less clearly associated with a specific attack on a specific target and more geared toward denying the attacker the ability to wage war in general.

Warfare has taught us again and again that active defenses and counterattacks are far more effective than passive defenses. Look at the Battle of Gettysburg in the American Civil War. Look at the Battle of the Bulge in World War II. Look at Leyte, Agincourt, and almost any piece of military history. Even in the animal kingdom, teeth and claws are a better defense than a hard shell or fast legs.

On the Internet, most people think of computer security in terms of passive defenses only. They believe that if they could only make their systems “hard” enough, they’d be safe. Security vendors reinforce this view, providing ever more intricate protection mechanisms for computers and networks. Even the work I’ve done, pointing out the limitations of prevention and extolling the virtues of detection and response, are still centered around passive defense. Part I of this essay was similarly limited: the ability to shift forces and knowledge of the terrain are both primarily associated with passive defense.

If we’re ever going to win the war against computer crime, we’re going to have to increasingly think more in terms of active defenses, and even counterattacks.

We’ve started to see some of this already. Intrusion detection systems and honeypots provide alarms that can alert defenders of an attack in progress. Managed Security Monitoring services can filter these alarms and provide expert response when a network is under attack. Vigilant, adaptive, relentless, expert intelligent network defense is far more effective than static security products. I said all of this in Part I of this essay.

But alarm systems, no matter how effective, are still primarily passive. They allow a defender to better survive an attack in progress, but they don’t put the attacker in danger. Right now, the only counterattack we have is prosecution. Putting criminals in jail is the best deterrent we have, and I am happy to see more of it. But prosecution can only happen after the fact.

One can imagine active defenses and counterattacks, but they are mostly in the realm of science fiction. What if, when an attacker broke into a network, his attack program were disabled? What if he could be sent a virus that destroys his computer? Or, at least, what if some third party collected an evidentiary chain that could prove his guilt in court?

There are non-technical considerations as well. In most countries, active defenses can be illegal. Private citizens can’t mine their backyards or booby-trap their front doors. In many countries, it is illegal for them to shoot a burglar breaking into their house. Active defenses are reserved for wartime, where there are no rules, or for the police, who have a state-sponsored monopoly on violence.

I worry about the vigilante-style cyber-justice that could arise from this kind of defense, but it is certainly something we should be thinking about. And it is definitely something that we should be researching.

Passive defense is far from useless, but is not the only form of defense we can use. In many cases, simple active defenses such as monitoring are both more effective and more cost effective than adding more passive defenses. “Fortress computer center” was a good model when every company had its own unconnected networks. In today’s world, where every network must be connected to the global network, it doesn’t work as well. If we are ever going to win the war against computer crime, we are going to have to emerge from our protective bunkers and actively engage the attacker.

Crypto-Gram Reprints

Computer Security: Will we Ever Learn?

Trusted Client Software

The IL*VEYOU Virus (Title bowdlerized to foil automatic e-mail traps.)

The Internationalization of Cryptography

The British discovery of public-key cryptography

The Futility of Digital Copy Prevention

Music, videos, books on the Internet! Freely available to anyone without paying! The entertainment industry sees services like Napster as the death of its business, and it’s using every technical and legal means possible to prevail against them. They want to implement widespread copy prevention of digital files, so that people can view or listen to content on their computer but can’t copy or distribute it.

Abstractly, it is an impossible task. All entertainment media on the Internet (like everything else on the Internet) is just bits: ones and zeros. Bits are inherently copyable, easily and repeatedly. If you have a digital file—text, music, video, or whatever—you can make as many copies of that file as you want, do whatever you want with the copies. This is a natural law of the digital world, and makes copying on the Internet different from copying Rolex watches or Louis Vuitton luggage.

What the entertainment industry is trying to do is to use technology to contradict that natural law. They want a practical way to make copying hard enough to save their existing business. But they are doomed to fail.

For these purposes, three kinds of people inhabit the Internet: average users, hackers, and professional pirates. Any security measure will work against the average users, who are at the mercy of their software. Hackers are more difficult to deter. Fifteen years of software copy protection has taught us that, with enough motivation, any copy protection scheme—even those based on hardware—can be broken. The professional pirate is even harder to deter; this is someone willing to spend considerable money breaking copy protection, cloning manuals and anti-counterfeiting tags, even building production plants to mass-produce pirated products. If he can make a profit selling the hacked software or stolen music, he will defeat the copy protection.

The entertainment industry knows all of this, and tries to build solutions that work against average users and most hackers. This fails because of a second natural law of the digital world: the ability of software to encapsulate skill. A safe that can keep out 99.9% of all burglars works, because the safe will rarely encounter a burglar with enough skill. But a copy protection scheme with similar characteristics will not, because that one-in-a-thousand hacker can encode his break into software and then distribute it. Then anyone, even an average user, can download the software and use it to defeat the copy protection scheme. This is what happened to the DVD industry’s Content Scrambling System (CSS). This is how computer games with defeated copy protection get distributed.

The entertainment industry is responding in two ways. First, it is trying to control the users’ computers. CSS is an encryption scheme, and protects DVDs by encrypting their contents. Breaks do not have to target the encryption. Since the software DVD player must decrypt the video stream in order to display it, the break attacked the video stream after decryption. This is the Achilles’ heel of all content protection schemes based on encryption: the display device must contain the decryption key in order to work.

The solution is to push the decryption out of the computer and into the video monitor and speakers. To see how this idea helps, think of a dedicated entertainment console: a VCR, a Sega game machine, a CD player. The user cannot run software on his CD player. Hence, a copy protection scheme built into the CD player is a lot harder to break. The entertainment industry is trying to turn your computer into an Internet Entertainment Console, where they, not you, have control over your hardware and software. The recently announced Copy Protection for Recordable Media has this as an end goal. Unfortunately, this only makes breaking the scheme harder, not impossible.

The industry’s second response is to enlist the legal system. Legislation, such as the Digital Millennium Copyright Act (DMCA), made it illegal to reverse-engineer copy protection schemes. Programs such as the one that broke CSS are illegal to write or distribute under the DMCA. This is failing because of a third natural law of the digital world: the lack of political boundaries. The DMCA is a U.S. law, and does not affect any of the hundreds of other countries on the Internet. And while similar laws could be passed in many countries, they would never have the global coverage it needs to be successful.

More legal maneuvering is in the works. The entertainment industry is now trying to pin liability on Internet service providers. The next logical step is to require all digital content to be registered, and to make recording and playback equipment without embedded copy protection illegal. All in an attempt to do the impossible: to make digital content uncopyable.

The end result will be failure. All digital copy protection schemes can be broken, and once they are, the breaks will be distributed…law or no law. Average users will be able to download these tools from Web sites that the laws have no jurisdiction over. Pirated digital content will be generally available on the Web. Everyone will have access.

The industry’s only solution is to accept the inevitable. Unrestricted distribution is a natural law of digital content, and those who figure out how to leverage that natural law will make money. There are many ways to make money other than charging for a scarce commodity. Radio and television are advertiser funded; there is no attempt to charge people for each program they watch. The BBC is funded by taxation. Many art projects are publicly funded, or funded by patronage. Stock data is free, but costs money if you want it immediately. Open source software is given away, but users pay for manuals and tech support: charging for the relationship. The Grateful Dead became a top-grossing band by allowing people to tape their concerts and give away recordings; they charged for performances. There are models based on subscription, government licensing, marketing tie-ins, and product placement.

Digital files cannot be made uncopyable, any more than water can be made not wet. The entertainment industry’s two-pronged offensive will have far-reaching effects—its enlistment of the legal system erodes fair use and necessitates increased surveillance, and its attempt to turn computers into an Internet Entertainment Platform destroys the very thing that makes computers so useful—but will fail in its intent. The Internet is not the death of copyright, any more than radio and television were. It’s just different. We need business models that respect the natural laws of the digital world instead of fighting them.

Similar sentiment about the death of the PC:


“Nihil tam munitum quod non expugnari pecuna possit.” So said Marcus Tullius Cicero, a Roman poet, statesman, philosopher and writer who is supposed to have lived 106-43 B.C. Translation: “No place is so strongly fortified that money could not capture it.” (I know this is not news, but it’s interesting.)

A bug in commercial PGP that allows an attacker to drop files to your disk that may then get executed (thanks to Windows .dll loading from current working directories).

An excellent article on the dangers of UCITA:

There is a security flaw in Alcatel DSL modems:
Normally, I wouldn’t even bother with this story. But Alcatel posted a MS Word file on their Web site about the problem and fix (which they’ve since removed). Unfortunately, the file saved deleted changes. The draft document is far more interesting than the real one. See some of the deleted comments here:

Microsoft responded to my article on the fake certificates in the previous Crypto-Gram:
Greg Guerin has rebutted Microsoft’s claims better than I could:
It turns out that the truth is way more complicated, but no more secure, than I had originally thought.

Remember the break last December? Here the CEO discusses what he would and wouldn’t do differently if faced with the situation again:

Anti-sniffing password management software. I’m not convinced this will work, but at least people are thinking about the problem. Shareware.

_Body of Secrets_ by James Bamford. This is his second book about the NSA, and it’s really good. I did a review for Salon:
Here’s another review from The New York Times:

CERT is charging companies to get early warnings about threats and vulnerabilities. On the one hand, it’s nice to see a little free enterprise here. On the other hand, isn’t CERT government-funded? But CERT advisories often appear long after other newsgroups report on vulnerabilities, so I don’t know how valuable this service really is.

Giga has released a report on the Managed Security Services space. It says nice things about Counterpane, but that’s almost beside the point. There has been a lot of confusion in the security services space, and the author nicely segments the businesses into six categories. He does a good job explaining what the different managed security services are, and which companies offer what services.

It’s hard to take this particular story seriously, but I have long predicted that insurance companies will start differentiating premiums based on what kind of networking hardware and software you use:

Impressive investigative work by the FBI. This is the kind of thing I like to see the FBI doing, rather than mucking about with surveillance tools like Carnivore.
Some disagree with me:

Years ago, ftp was how you shared files between computers. There are still vulnerabilities associated with this service.

A major legal battle is looming, as the RIAA tries to suppress Princeton security research into its digital watermarks, citing secrecy provisions of the DMCA:
A preliminary version of the actual paper, and assorted correspondence:
The site reported over 50,000 visits to the paper within 24 hours of its posting.
The RIAA changes its tune:

Don’t forget mundane security risks. The British Ministry of Defense has lost 205 laptops in the past four years.

An e-mail was recently sent to Amazon associates, inviting them to visit a non-Amazon Web site and complete a questionnaire. The e-mail purported to come from, but was actually sent from an entirely different domain <>. When I asked Amazon whether they were being spoofed, they told me the survey was legitimate. Are they trying to train their customers to respond to unverified impersonations?

Argus boasted that their secure operating system couldn’t be hacked, and sponsored a $50K contest. It was hacked. The story of how it happened has a moral for everyone: security is only as strong as the weakest link, and if you’re not monitoring your security in real time you need to constantly make sure all the links are strong.
Someone else plans on a $1M hacking contest.

Gene Spafford makes much the same points I do about the future of computer security: it’s going to get worse, not better.

There have been zillions of articles on this “May Day Cyberwar.” Supposedly, the Chinese are attacking the U.S. in retaliation for our lousy foreign relations policies.
I believe this is nothing but hacker fantasy and media hype. I don’t see hackers with political motivations taking up arms; I see hackers with no motivations donning a cloak of politics to justify their actions. I also see the media turning this into a much bigger deal than reality.

People are the weakest link in security:

U.S. “national security” surveillance is on the rise:

Cyber-thriller screenplay:

Comments on NIST’s AES FIPS are due by May 29th. This isn’t the time to suggest alternate algorithms, but it is time to comment on the details of the standard.

The Dutch government is forcing trusted third parties to use key escrow.

Another semantic attack. A fake BBC Web page was circulating (without the caveat at the top), and the British newspapers fell for it.
The fake Web page (with a disclaimer on the top):

Microsoft and the Window of Vulnerability

In many of my speeches, I talk about a “Window of Vulnerability.” When a security vulnerability exists in a product and no one knows about it, there is very little danger. But this state of security is fragile. As soon as someone discovers the vulnerability, the danger increases. If we’re lucky, the discoverer is a good guy who does not exploit the vulnerability for personal gain. Eventually word of the vulnerability gets out, and the danger increases.

This sounds just like the real world, but cyberspace has a crucial difference. If I knew how to break into a certain kind of ATM, or hot-wire a certain make of car, or pick a certain model of lock, I could teach someone. The person I taught would then know how, and he could teach others. But it’s a skill, and skills take time to teach. Cyberspace is different because skill can be encapsulated into software. If I knew how to break into Microsoft’s IIS 5.0, I could turn my knowledge into an exploit and distribute it on the net. Then, hundreds of thousands of “script kiddies”—with no skill whatsoever—could use my exploit to break into IIS 5.0. The propagation characteristics of virtual vulnerabilities are very different than physical vulnerabilities.

We’re seeing this happen right now with an IIS 5.0 vulnerability. It was discovered by a company called eEye Digital Security, which was nice enough to warn Microsoft and give them time to create a patch. Then, Microsoft and eEye announced both the vulnerability and the availability of a patch. A few days later, someone wrote an exploit. As the exploit made its way through the hacker community, and continues to do so, more and more IIS installations are being broken into.

The press regularly writes the story like this. First, vulnerability discovered and we’re all in danger. Then, vulnerability patched and we’re all safe again. What they forget is that patches don’t work unless they’re installed. And more and more often, people don’t install patches. I predict that years from now, Web sites will still be broken into because of this vulnerability.

So here’s the million-dollar question: Is eEye Digital Security part of the solution, or is it part of the problem? eEye’s own legal disclaimer implies that even they’re not sure: “In no event shall the author be liable for any damages whatsoever arising out of or in connection with the use or spread of this information.”

Microsoft IIS vulnerability:
eEye Digital Security’s announcement:
Microsoft security advisory and patch information:
Exploit published:

Schneier’s essay “Closing the The Window of Exposure”:
The fallacy of installing patches:

Counterpane Internet Security News

There have been an enormous number of exciting things going on at Counterpane. I can’t talk about any of it yet, because we’re still working on press releases. We acquired SDII, a small consulting company.
More news next month.

Articles on Counterpane have appeared in The New York Times and The Economist:

eWeek reported on Schneier’s talk at the RSA Conference last month:

Bruce Schneier is speaking at ISSA events in New York (May 17), Palo Alto (Jun 6), and Denver (Jun 14):

Schneier is speaking at the Trema World Forum in Monaco on May 30:

_Secrets and Lies_ won a “Jolt” award from Software Development magazine:

And Counterpane is still hiring:

Security Standards

Andrew Tanenbaum once quipped that the great thing about standards is that there are so many to choose from. Despite numerous efforts over the years to develop comprehensive computer security standards, it’s a goal that remains elusive at best.

It all started with the Orange Book. As far back as 1985, the U.S. government attempted to establish a general method for evaluating security requirements. This resulted in the “Orange Book,” the colloquial name for the U.S. Department of Defense Trusted Computer System Evaluation Criteria. The Orange Book gave computer manufacturers a way to measure the security of their systems and offered a method of classifying different levels of computer security.

The goal was to aid government procurement, but it also held the promise of benefiting the entire industry as well. That never came to pass, primarily because certification testing was expensive and controlled by only a few labs, and the resulting designations weren’t well-suited to the civilian marketplace’s needs.

There have been other efforts over the years to codify security, but they were unsuccessful. Now, several industries are rallying around the Common Criteria, an ISO standard (15408, version 2.1) that provides a catalog of security features such as confidentiality and authentication. Companies and industries using this document are expected to include these concepts in a more specific “protection profile,” which is basically a statement of security requirements.

Then, individual products can be tested against that profile. For example, a smart card could be tested against a protection profile with such attributes as resistance to cloning, security of protocols and protection against physical reverse engineering, and a firewall could have a different protection profile that includes attributes related to its security and functionality.

It’s a great idea, and puts more meat on the bone than past efforts. But don’t expect it to work except in a few isolated areas. The problem is that these standards are too general. They won’t tell you how to configure your CheckPoint firewall, or what security settings to run on Windows 2000. It’s not a shortcoming in the standards; it’s just not feasible to document an infinite number of scenarios.

Consider something truly quantitative: say, a configuration guide on the best way to secure Red Hat Linux 6.0. It could be an excellent standard, but it will probably be obsolete in a few weeks. It will certainly have to be revised for version 6.1. And it can’t possibly help you configure Solaris version 3.2, let alone Windows NT SP 4.0.

On the other hand, some standards can be too specific, making it almost impossible to test a general system. Remember when Windows NT received the Orange Book’s C2 security rating? The rating was only good for a specific configuration of Windows, one unconnected to the network and without any removable media. What about a rating for the overall security of Windows NT? Forget about it!

The bottom line is that while these standards can be very useful for certain applications, they aren’t useful gauging enterprise security in general. The Common Criteria is a great document, and companies like Visa are putting a lot of effort to turn it into something that they can use for their own purposes. The credit card company is currently using the document to specify security levels of hardware and software. But that’s only a special case; no one else can take what Visa did and make use of it.

I have long joked that given any general security standard, I could design a product that 1) met the standard, and 2) was still insecure. Given this truism, it’s no wonder that these standards don’t find much utility in the commercial world. And it’s no wonder why there are so many standards to choose from.

Common Criteria:

NSA’s Rainbow Series, including the Orange Book:

There are configuration guides that are designed to help you with specific products. This SANS Windows NT guide is an excellent example:
So is Phil Cox’s Windows 2000 guide:
<> [link moved to]

Safe Personal Computing

I am regularly asked what the average Internet user can do to ensure his security. My first answer is usually “Nothing; you’re screwed.” But it’s really more complicated than that.

Against the government there’s nothing you can do. The power imbalance is just too great. Even if you use the world’s best encryption, the police can install a keyboard sniffer while you’re out. (If you’re paranoid enough to sleep with your gun and laptop under your pillow, this article is not written for you.) Even big corporations are difficult to defend against. If they have your credit card number, for example, there’s probably no way to make them forget it.

But there are some things you can do to increase your security on the Internet. None of these are perfect; none of these are foolproof. If the secret police wants to target your data or your communications, none of these will stop them. But they’re all good network hygiene, and they’ll make you a more difficult target than the computer next door.

1. Passwords. You can’t memorize good enough passwords any more, so don’t bother. Create long random passwords, and write them down. Store them in your wallet, or in a program like Password Safe. Guard them as you would your cash. Don’t let Web browsers store passwords for you. Don’t transmit passwords (or PINs) in unencrypted e-mail and Web forms. Assume that all PINs can be easily broken, and plan accordingly.

2. Antivirus software. Use it. Download and install the updates every two weeks, and whenever you read about a new virus in the media. Some antivirus products automatically check for updates.

3. Personal firewall software. Use it. There’s usually no reason to allow any incoming connections from anybody.

4. E-mail. Delete spam without reading it. Don’t open, and immediately delete, messages with file attachments unless you know what they contain. Don’t open, and immediately delete, cartoons, videos, and similar “good for a laugh” files forwarded by your well-meaning friends. Turn off HTML mail. Don’t use Outlook or Outlook Express. If you must use Microsoft Office, enable macro virus protection; in Office 2000, turn the security level to “high” and don’t trust any sources unless you have to. If you’re using Windows, turn off the “hide file extensions for known file types” option; it lets Trojan horses masquerade as other types of files. Uninstall the Windows Scripting Host if you can get along without it. If you can’t, at least change your file associations so that script files aren’t automatically sent to the Scripting Host if you double-click them.

5. Web sites. SSL does not provide any assurance that the vendor is trustworthy or that their database of customer information is secure. Think before you do business with a Web site. Limit financial and personal data you send to Web sites; don’t give out information unless you see a value to you. If you don’t want to give out personal information, lie. Opt out of marketing notices. If the Web site gives you the option of not storing your information for later use, take it.

6. Browsing. Limit use of cookies and applets to those few sites that provide services you need. Regularly clean out your cookie and temp folders (I have a batch file that does this every time I boot.) If at all possible, don’t use Microsoft Internet Explorer.

7. Applications. Limit the applications on your machine. If you don’t need it, don’t install it. If you no longer need it, uninstall it. If you need it, regularly check for updates and install them.

8. Backups. Back up regularly. Back up to disk, tape, or CD-ROM. Store at least one set of backups off-site (a safe-deposit box is a good place) and at least one set on-site. Remember to destroy old backups; physically destroy CD-R disks.

9. Laptop security. Keep your laptop with you at all times when not at home; think of it as you would a wallet or purse. Regularly purge unneeded data files from your laptop. The same goes for palm computers; people tend to keep even more personal data, including passwords and PINs, on them than on laptops.

10. Encryption. Install an e-mail and file encryptor (like PGP). Encrypting all your e-mail is unrealistic, but some mail is too sensitive to send in the clear. Similarly, some files on your hard drive are too sensitive to leave unencrypted.

11. General. Turn off the computer when you’re not using it, especially if you have an “always on” Internet connection. If possible, don’t use Microsoft Windows.

Honestly, this is hard work. Even I can’t say that I diligently follow my own advice. But I do mostly, and that’s probably good enough. And “probably good enough” is about the best you can do these days.

Comments from Readers

From: David Wallace <david.wallace>
Subject: Military History and Computer Security

I was taken aback by your assertion that a burglar alarm works because “the attacker doesn’t know they’re there.” After all, “true victory consists of breaking the enemy’s will without fighting.” The first line of defense is deterrence, the number one reason for installing a burglar alarm. Security starts with making yourself a more difficult target. Hence the “Premises protected by” stickers in windows and “Alarm” signs in front yards. They encourage a potential attacker to pick another, less heavily defended, target. In fact, the target may be completely undefended, protected only by signage purchased at a hardware or department store.

The Internet makes deterrence a little more dicey. First off, the alarm is necessary, but the “alarm” sign is impractical. It is a potential “red cape” waved at a hacking “bull.” It may also tip the defender’s hand by revealing his defenses. In the physical realm there are a wide variety of systems and sensors to deploy to “measure.” In the virtual, there are fewer, they are less easily understood, and harder to install and configure.

Once deterrence fails, detection becomes key. In the physical world, the alarm system monitors a variety of metrics to evaluate defensive posture (system armed/unarmed), readiness to respond (sensor operational/deactivated), and violations of its sensors (heat, motion, noise, moisture, or sensor loss). The Internet alarm performs the same functions, and performs them in much the same way.

The next step in deterrence is the concept of “unacceptable losses”. Here the two worlds both converge and diverge. They converge on the definition of unacceptable losses. On both the physical and logical plane unacceptable losses include arrest, conviction, fine, and/or imprisonment. They diverge in the likelihood of suffering unacceptable losses.

As you note in _Secrets and Lies_, in physical security, the attacker must be physically present, rendering him not only detectable, but visible, and apprehend able. The Internet removes that risk from the attacker, allowing him to strike remotely and in relative anonymity.

Once attacked, there are two phases to the defense: Repel and counterattack. In the physical world, once an attacker is repelled, you follow up with counterattack. Repelling the attack is accomplished by holding ground and buying time while the resources needed to stop the attack are marshalled and committed (amateurs debate tactics, professional soldiers argue logistics).

Counterattack is accomplished by understanding the attacker’s objective and the resources he has committed to the attack. The defender manipulates these variables to expose vulnerabilities in the attacker’s position which can be exploited. These can weaken the enemy, forestall his attack, and potentially force his retreat. If retreat can be forced, it can be followed up with pursuit, further weakening the attacker, deterring future aggression, and potentially reducing the attacker’s resources below the level necessary to support another assault.

Unfortunately, counterattack and pursuit do not transfer well to the virtual battlefield. About the only option is to repel. The logical version of counterattack is limited to prosecution, which proves difficult when attacks occur across state and national boundaries. Even when prosecution does occur, it is hampered by poor forensics, poor laws, and general ignorance within the court system (See the judge in the Mittnick trial).

So what can you do to defend? Roll deterrence into your defense. Monitor. REVIEW THE LOGS! Have an incident response plan. Partner with law enforcement and a professional forensics team. Be prepared to go public when attacked. Aggressively prosecute intruders whenever possible. Develop a reputation as a target to stay away from.

From: Henry Spencer <henry>
Subject: Military History and Computer Security

I would argue that there’s a third issue, more important on the military side although it’s not clear that there is any useful Internet analogy. Another old military axiom: “the attacker must vanquish; the defender need only survive.”

The defender’s biggest advantage is that the attack has to make progress to succeed, and the defense doesn’t. This puts the attacker out in the open, moving forward, while the defender is stationary and under cover—less visible, better protected, and much more easily connected to communications and supply lines.

This shows, for example, in a traditional distinction between two types of hand grenades: offensive and defensive. An offensive grenade has a rather limited lethal radius, because it’s meant to be used by attackers, who may be on the move or behind poor cover; in particular, it relies more on blast than on fragmentation. A defensive grenade is designed to be lethal over the widest possible area, for use by people who are safely ensconced behind solid cover and may be (locally) badly outnumbered. (I am not sure this distinction is still made nowadays, since even defensive forces now tend to emphasize mobility, but at one time it was taken quite seriously.)

From: “Gerard Joseph” <gerard>
Subject: Military History and Computer Security

I keep thinking about the apportionment of blame between the innocent defender and the guilty attacker. Presumably, a bank robber would still be charged and found guilty even if one night the bank completely forgot to lock its doors or set its alarms. But in that case I’m sure the bank would be held partly responsible for the attack. If someone takes a shot at me while I’m ambling on the street, then he will always be guilty, even though I might have been negligent in walking on that particular street at that particular time. It seems that in all cases there develops, over time and in accordance with local norms and experience, a state of equilibrium between the rate of crime and the level of defenses that are customarily implemented to thwart criminal acts. Ideally, this state represents an optimal balance between the level of crime and the cost of relevant defensive measures. A criminal who succeeds in spite of those defenses is more readily seen to be guilty, while a victim who falls short in implementing accepted levels of defense is less readily seen to be innocent. But in no case does the victim’s negligence excuse or justify the crime, nor does the criminal’s ability to overcome your defenses excuse or justify their absence.

I think as far as the Internet is concerned, we are groping towards the defining equilibrium between crime and defense. Right now, there is a set of protective measures whose omission would certainly represent culpability on the part of a defender, and there is a set of attacks whose commission would certainly represent a crime (whether legally recognized or not) on the part of the attacker. But in between there is a grey area of defenses and attacks that lack categorical classification. To date, though, I think we’ve been too lenient on both complacent defenders and aggressive attackers. That must and surely will change. A starting point would be for the media to stop interviewing hackers as if they were just ordinary community-minded citizens.

From: Stephen Tye <StephenT>
Subject: e-mail filter idiocy

I have read your article and I can understand your annoyance at having your e-mail blocked for containing the unrelated words “blow” and “job”. I admit the sample text censor scripts that we provided in MailMarshal version 3.3 have a couple of anomalies like this that would false trigger. We have done a lot of work on our sample text censor scripts for the next version release to improve them and minimize false triggers.

MailMarshal is a tool to allow companies to apply corporate policy to their e-mail. Technically MailMarshal did exactly what it was told to do, which was to block e-mails with the words blow and job in them. In this case it was the script that was at fault, not the product.

Depending on how the company has set up our product to match their corporate guidelines, it is highly likely that the intended recipient of your e-mail also received a notification e-mail informing them that your e-mail did not arrive. The e-mail you sent would have most likely been quarantined and could have been easily released by the administrator. The line “blow and job” could have then been removed from the text censor script and the problem would never occur again.

If it is the organization’s policy to block any e-mails which contain the words “IL*VEYOU” in the subject, then that is their choice and MailMarshal will allow them to enforce that policy. We normally only suggest using a text censor script in this way when there is a virus alert and you would like implement some protection until you can get your antivirus product updated. Otherwise we find scanning e-mails with an antivirus product and implementing rules that block e-mails which contain EXE or VBS attachments (which normally have no business use for end users) an effective protection against e-mail borne viruses.

As you well know, security is process, not product. MailMarshal is a tool that allows you to apply that process. It will only action what it has been told to do.

CRYPTO-GRAM is a free monthly newsletter providing summaries, analyses, insights, and commentaries on computer security and cryptography.

To subscribe, visit <> or send a blank message to To unsubscribe, visit <>. Back issues are available on <>.

Please feel free to forward CRYPTO-GRAM to colleagues and friends who will find it valuable. Permission is granted to reprint CRYPTO-GRAM, as long as it is reprinted in its entirety.

CRYPTO-GRAM is written by Bruce Schneier. Schneier is founder and CTO of Counterpane Internet Security Inc., the author of “Secrets and Lies” and “Applied Cryptography,” and an inventor of the Blowfish, Twofish, and Yarrow algorithms. He served on the board of the International Association for Cryptologic Research, EPIC, and VTW. He is a frequent writer and lecturer on computer security and cryptography.

Counterpane Internet Security, Inc. is a venture-funded company bringing innovative managed security solutions to the enterprise.


Sidebar photo of Bruce Schneier by Joe MacInnis.