That 5th Amendment computer case seems to becoming more and more important as time goes on. Here is a recent article from ABC News on the subject.


Lock Out the FBI for $200
Experts: Encryption Is Easy Way to Keep Prying Eyes Out of Your Hard Drive

By ASHLEY PHILLIPS and SCOTT MICHELS, ABC News
Feb. 11, 2008 �


The widespread belief that any database, hard drive or electronic device can be hacked was disproved when a man accused of having child pornography on his computer managed to keep federal authorities out of his hard drive for more than a year -- for the price of an average cell phone.

That computer protection used by the suspect is easy to obtain, even common on most computers, and, according to security experts, is almost impossible to breach, even for the FBI.

On Dec. 17, 2006, Sebastien Boucher was stopped by border patrol inspectors while crossing from Canada into Vermont. An inspector found a laptop in his car, which Boucher admitted belonged to him, according to an affidavit from an Immigration and Customs Enforcement agent.

After Boucher gave the agents access to his computer, they saw videos and file names that appeared to show pornography involving pre-teens, including one labeled "Two-year-old being raped during diaper change." Boucher, a Canadian citizen who is a lawful U.S. permanent resident, said he didn't know if his computer had child pornography because he could not check his temporary Internet files, the affidavit says. He was arrested and charged with transportation of child pornography, a felony that carries up to 20 years in prison.

But after Boucher's arrest, an investigator from the Vermont Department of Corrections was unable to access the images on Boucher's computer, which were stored in an encrypted drive called drive Z.

For more than a year, the government has not been able to see what is in drive Z, which is protected by an encryption program that is sold under the name Pretty Good Privacy, according to court records.

Pretty Good Privacy, which is more commonly known as PGP, is an industry standard of hard-drive encryption and email encryption, according to experts. Encryption is a complex, password-protected method of keeping information, hard drives, devices - almost anything - private.

"If you hand me someone's normal laptop, it is relatively easy to bypass passwords. All you have to do is rip out the hard drive out and put it into a different computer," said Charles Miller, a principal security analyst at Independent Security Evaluators and former employee of the National Security Agency. "PGP is full-disk encryption, which means the entire disk is encrypted and the only way in is to know the password. The program makes a key and that key is a password, without it you can't get into to the drive."

A desktop PC version of PGP is available for less than $200, and open-source (read: free) versions, sometimes called GPG, can be found online. Similar encryption services are also available in standard operating systems on PCs and Macs. Consumers often don't use them, however, because if they lose their password, there's no way to retrieve the protected information.

"People can't snoop because of strong encryption... It is similar to what protects your information and money at a bank," Miller said.

The software has proven to be instrumental in Boucher's case.

Secret Service Agent Matthew Fasvlo testified at a court hearing in 2007 that it is "nearly impossible" to access the encrypted files without the password.

"There are no 'back doors' or secret entrances to access the files. The only way to get access without the password is to use an automated system which repeatedly guesses passwords," Magistrate Judge Jerome Niedermeier, who was presiding over the case, wrote. "According to the government, the process to unlock drive Z could take years."

Jonathon Giffin, an assistant computer science professor at Georgia Tech, said without the password there was only one way to get into the computer: with "brute force."

"They start trying all possible passwords, hoping that they have passwords that you use," Giffin said. "The expected time it would take is years, decades, unless you have extremely powerful computers."

Even the FBI doesn't have that kind of computing power, according to Giffin.

"The FBI probably does not. The NSA probably does," he speculated. "That's really one of the NSA's jobs - to develop cryptosystems for our military as well as to crack the cryptosystems of other governments."

The government subpoenaed Boucher to try to force him to type in his password and give the government access to the computer.

In November, Niedermeier ruled that forcing Boucher to enter his password would violate his Fifth Amendment right against self-incrimination.

"If Boucher does know the password, he would be faced with the forbidden trilemma; incriminate himself, lie under oath, or find himself in contempt of court," the judge wrote.

In the 1990s, according to Virgil Gligor, an electrical and computer engineering professor at Carnegie Mellon University, the U.S. government attempted to prevent situations like Boucher's.

"In the late '90s, there was a government initiative, in which the government required any encryption mechanism to save keys [or passwords] and give keys to the government," Gligor said. "It failed."

Even since then, encryption programs have only improved.

"The quality of encryption that we have nowadays is actually ...very good," Gilgor said.

Without that type of regulation proposed a decade ago, the government has been left with one option: to appeal the court's decision.

Russell Goldman contributed to this report.


Copyright � 2008 ABC News Internet Ventures

Don't be crook, and you can keep your $200.

Don't underestimate law enforcement. The reason there are aren't more "cracked" passwords showing up in the case law probably has as much to do with DOJ legal strategy as it has to do with an inability to crack them. Actually, I think this is what is at the root of the recent spate of "hard drive/encryption" reports and cases we are seeing.

The "Law Enforcement" community actually has a very effective password cracking system. If you know any feds, ask them what their office computers do in "background" mode. Literally tens of thousands of federal LE computers are components of a DNA (distributed network attack) system often budgeted as "Project Safe-Cracker". The program reportedly started nationwide with Secret Service PCs but has been systematically expanded. It operates like the "SETI at home" project, if you recall that. A central computer takes a very hard problem (like cracking a password) and 'parcels' the problem into thousands of lesser problems, sending them out over the network to be worked on by PCs, university mainframes, etc. As they solve their bits of the problem they communicate back to the central controller. The software runs in "background" and the agencies don't normally even tell the users about it being there. Just ask your fed IT guy, he'll fill you in.

Here is the real "magic," however. Humans are terrible at making, let alone remembering, truly 'random' passwords. Virtually every password has "meaning" to the person who makes/uses it. (for example, you use "your wife's initials + wedding year" for a password). Now, if someone had "complete knowledge" of you they could come up with a "small" list of a few billion (which is trivial by todays standards) potential password combinations, one of which would be very likely to be the key/password.

How do you get such "total knowledge"? The most efficient way is normally to get a copy of the person's hard drive. You take all the information on the hard drive, then take copies of all the web pages ever visited, etc. etc. and use it as a "data set" to be filtered by very sophisticated algorithms into sets of likely passwords. (that is for relatively sophisticated apps like PGP, less secure encryption can often be cracked by mere 'brute force'.)

You can usually *exponentially* reduce the amount of time it takes to crack a password by using this type of technique. What you need is to get your hands on that danged "data set". This where search warrants for hard-drives, and often fuzzy "search and seizure" law, come in (particularly border searches). And the feds love fuzzy. I'd bet dollars to donut the DOJ is waiting for the "perfect precedent" to get wide latitude to keep/use *all* data recovered by search/seizures for purposes such as password cracking.

If you need a password cracked, feel free to ask the DHS if you can use "safecracker." Unless your perp speaks Arabic they will probably say no, if only because they don't want case law to tangle things up.

But you could probably get a federal grant and start a state level program on Texas state computers.... (though you'd probably have to fight the feds tooth-and-nail for time on university supercomputers).

If that's the case, then why haven't the feds been able to crack this guy's password after MORE than one year of effort?

Below is a simple explanation as to why the Feds will not be able to crack the password that i found on wikipedia.


The resources required for a brute force attack scale exponentially with increasing key size, not linearly. Doubling key size does not double the required number of operations, but rather squares the number of required operations. Thus, although 56 bit keys, such as those used by the obsolete Data Encryption Standard (DES) are now quite practical to attack by brute force, this is not true of much longer keys, such as those used by the more modern Advanced Encryption Standard (AES), which uses keys of at least 128 bits in length.

There is a physical argument that a 128 bit key is secure against brute force attack. The so-called Von Neumann-Landauer Limit implied by the laws of physics sets a lower limit on the energy required to perform a computation of ln(2)kT per bit erased in a computation, where T is the temperature of the computing device in kelvin, k is the Boltzmann constant, and the natural logarithm of 2 is about .693. No irreversible computing device can use less energy than this, even in principle.

Thus, in order to simply flip through the possible values for a 128-bit key (ignoring doing the actual computing to check it), one would need a device consuming at a minimum 10 gigawatts (about the equivalent of eight large, dedicated nuclear reactors) running continuously for 100 years. The full actual computation - checking each key to see if you have found a solution - would consume many times this amount.

However, this argument assumes that the register values are changed using conventional set and clear operations which inevitably generate entropy. It has been shown that computational hardware can be designed not to encounter this theoretical obstruction: see reversible computing. It should be pointed out that such computers have not yet been constructed.

The amount of time required to break a 128 bit key is also daunting. Each of the 21 times 10 to the 28th power possibilities must be checked. This is an enormous number, 340,282,366,920,938,463,463,374,607,431,768,211,456 in decimal. A device that could check a billion billion keys (10 to the 18th power) per second would still require about 10 to the 13th power years to exhaust the key space. This is longer than the age of the universe, which is about 13,000,000,000 (1.3 times 10 to the 10th power) years.

AES permits the use of 256 bit keys. Breaking a 256 bit key by brute force requires 21 times 10 to the 28th power time more computational power than a 128 bit key. A device that could check a billion billion (10 to the 18th) AES keys per second would require about 10 to the 51st power years to exhaust the 256 bit key space.

Hence, 128 bit keys are impractical to attack by brute force methods using current technology and resources, and 256 bit keys are not likely to be broken by brute force methods using any obvious future technology.

[This message was edited by RTC on 02-12-08 at .]

Don't the feds have that thing where you just swipe a card and the red digital display scrolls past various digital represnations of Sting, Stewart Copeland, and Andy Summers until all seven numbers fall into place?

Or do they only sell that at cat-burglar/safe-cracker stores?

[This message was edited by David Newell on 02-12-08 at .]

To begin with, the very weakness of cipher systems that use human-generated passwords is the fact that humans will rarely if ever actually generate passwords (i.e. keys) that come anywhere near filling the keyspace. A truly random'128 bit' key/password might look like this:
4GqT5hOEHKveNnVP8TcUskCY1rzOpkETQnEKyTpbrmGgyJuFLCnURGohKnMb6ZE

Now, how often do you actually use a string like that as a password? Basically, using XXX bit encryption does not necessarily help much if people use passwords like "fredlikespizza" or "ibornjuly". Even throwing in the odd 'random' number or special character will usually not radically alter the amount of keyspace that is actually likely to be utilized by a human user. As even Wikepedia cites: "In one survery of MySpace passwords which had been phished, 3.8 percent of passwords were a single word found in a dictionary, and another 12 percent were a word plus a final digit; two-thirds of the time that digit was 1." A lot of criminals are dumb, or just smart enough to not do a good job.

Secondly, while more secure ciphers like PGP are availabe, they are certainly not the rule. Many ciphers are very much simpler, for example the default system used by Word and Excel.
Check out a web page like this one:
http://lasecwww.epfl.ch/~oechslin/projects/ophcrack/

thirdly, computationally difficult attacks don't necessarily have to be either brute force or involve human engineering. There are so called "side channel attacks" that can benefit from DNA and related methodologies. For example, differential power analysis data tables.

Why has this guys password not been cracked in a year? It could be an exception (there are crooks who do exceptionally good jobs -- fortunately they are the exception). Passwords are routinely cracked, but when they are you don't need to compel the defendant to divulge the password and thus make newsworthy legal precedent. Show them the decrypted files and take your plea, usually.

On a final note, this case may not have gotten "full treatment" for policy reasons. You'd have to ask DHS what their policy is on the use of Safecracker. I would imagine terror investigations go to the top of the heap. Some ciphers/passwords are simply not considered worth the time by system managers. As always there are trade-offs.

Password cracking, like just about any forensic tool, may not work in every single case. It can, however, work in a very useful number of them (if not the vast majority). Using password 'guessing' algorithms is made vastly easier with access to a lot of data about a person, and the easiest place to get that data is that person's hard drive. How to do all this within the constraints of Constitutional protections is an open question.

When I did it, we didn't need no steeenkin' constitutional constraints -- fortunately the commies were foreigners.

My guess as to why the password has not been cracked, without knowing more, is that they have not even tried. FBI and NSA are probably using their computer time on national security stuff. The Feds are using a normal tactic to get around the encryption--legal compulsion of the key. Maybe its just me, but I don't see the difference between compelling a voice sample or a bank record or the encryption key. If I can make someone give a voice sample or fingerprint, I ought to be able to make them give me a word. Not as testimony in court, but as mere evidence. He can physically refuse, but the law attaches a penalty to that refusal.

Not really the same. In this case, he would have to state that he knows the password and then give it. Everyone has fingerprints... you do not have to admit you have one first.

Nope, I'm not seeing it. You're not asking the guy to admit he has child porn, or that he has porn. You are compelling him to provide a piece of historical information--what is your password? What is your account number? Because it is compelled, his particular answer cannot be used in that testimonial form, but the prosecution gets the benefit of it. The government already knows he has a password, because he already used it once. The government already knows the computer contains illegal pornography because a government agent saw it after the suspect used his password. Now, if he is compelled to produce the password, and the government finds on the laptop the location of Jimmy Hoffa's body because this fellow is the one who put him in the south end zone of the Meadowlands, he will be ok. Look at US v. Hubbell, 530 US 27 (2000).

On the face of it, this case is simple - a person trying to carry pornographic images of children across an international border. The catch is whether the courts can compel him to voluntarily abandon his rights under the Fifth Amendment.


Orin Kerr, a law professor and computer crime expert at George Washington University, said the distinction that favors the government in Boucher's case is that he initially cooperated and let the agent look at some of the laptop's contents.

"The government can't make you give up your encryption password in most cases. But if you tell them you have a password and that it unlocks that computer, then at that point you no longer have the privilege," he said.

Lee Tien, the attorney with the Electronic Frontier Foundation, said a person's right to keep a password secret is a linchpin of the digital age.

Encryption is "really the only way you can secure information against prying eyes," he said. "If it's too easy to compel people to produce their crypto keys, it's not much of a protection."

Another point to consider is that if Magistrate Niedermeier's ruling is allowed to stand, the result could be "dangerous" for law enforcement. According to Mark Rasch, a privacy and technology expert with FTI Consulting and former federal prosecutor, "If it stands, it means that if you encrypt your documents, the government cannot force you to decrypt them. So you're going to see drug dealers and pedophiles encrypting their documents, secure in the knowledge that the police can't get at them."

At the end of the day, we have to consider some truly gripping questions. First, we have to recognize that the rule of law is the rule of law for all. If you fall into U.S. jurisdiction, you are subject to whatever ruling is mandated as "law" on this question.

If we say that by encrypting the files, the individual had a reason to believe that the information should be private, is it okay to say that when the individual is a suspected terrorist? Or pornographer? Or senior business official? Is one better than another?

If we decide that a person's Fifth Amendment rights are inconsequential, are they always inconsequential? Are we then compelled to self-implicate? Where does the Fifth fit in?

This brings us to some crucial underlying questions - Where is the line, and HOW do we draw it? Or is the question HOW do we define the line? Can law truly BE case-by-case?

While the first person to test existing law has an unsavory rationale, I have to ask - would you care more if the files under fire were private business documents? Would you feel differently if they were personal documents between you and your significant other? Would you feel differently if they were personal documents between you and your terrorist cell?

How do you think these documents should be handled? Because of their very nature, it seems as if each incident should be considered on its own merits, but how do we define the supporting law?

This isn't about one guy with questionable content on his laptop. It is much deeper, and the impacts of the answers are far-reaching.

Can you excuse pornography, even child pornography, to keep your business safe? How about your country? How do we define the boundaries, and what is our message to the law makers? How about to law breakers?

I dunno. Seems like overthinking and overdramatizing. There's a body of law out there relevant to this, and we just need to see what the Supreme Court says. If they say that a compelled password is testimonial for the purposes of the 5th Amendment, that would be difficult for law enforcement. If you think about it, if we get something unencrypted, we still have to come up with ways to authenticate it, prove possession, prove knowledge, etc. Mere possession of the encryption key doesn't do that. I just don't see why the Court would extend the protection to a person like this.

From a practical point of view, the crooks can still protect themselves by refusing to provide the key and taking their six months in the cooler for contempt. Probably less time than they would get for the 10,000 images they have hoarded on their portable porn machine.

What if a person honestly can't recall a password?
Or better yet, what if the file is not encrypted intelligence at all, but simply random numbers? A truly good cipher system will produce output that looks like random. If you have a file that apppears to be "nothing but" random bytes, how do you know if there ever was a password? (sophisticated defensive strategies often involve "salting", inserting random data, to throw off attacks). For example, 'file scrubber' software often re-writes unwanted files with random data. There is no password/key and no simple way to get the data back. There may be no way to tell 'scrubbed' files from 'encrypted' ones.

Admitting that you know a password could be confirming the very existence of the underlying content, let alone that you had access to it.

Suppose a system has 2 passwords: one unlocks the system, the other deletes or destroys it.

Well, if you destroy something's value as evidence, we call that tampering. It is a third degree felony, punishable by 2-10 years in prison and a fine of up to $10,000. Tex. Penal Code Sec. 37.09. Self destructing the data would be tampering.

Maybe the solution would be to amend the tampering statute such that failure to obey a grand jury subpoena for an encryption key constitutes the offense of tampering. Heck, maybe the defendant's actions in this case constitute the offense of tampering already, because the encryption "conceals . . . a thing with the intent to impair its . . . legibility, or availability as evidence . . ." If it is encrypted, you can't read it, so it is not legible, and it is not available as evidence. I like that.

I was thinking more like James Bond using something easy for the self-destruct password... so that his captors delete it for him. Sort of a data trap or something.

quote:

---

If it is encrypted, you can't read it, so it is not legible, and it is not available as evidence. I like that.

---

HA! That IS very clever but by extending that logic everything stored on every computer disk is encrypted. People don't read bits and bytes and even if you did it would just be a bunch of 1s and 0s. It only means something if you know how to "decrypt" it. Nobody reads the raw data. Even computer professionals only examine it at this level when diagnosing a problem.

This is the phrase "Hello TDCAA" in ASCII binary:

01001000011001010110110001101100011011110010000001010100010001000100001101000001
01000001

That is 11 bytes long. I have a low end $99 digital camera and the default image size is 1MB so its on the order of 1000000 bytes long. In other words, impossible to read without a computer to decrypt it into an image.

Yes, but it is readable. Data stored in binary is encoded, but all computers of the proper type know the key. Encrypted data can only be encrypted with the proper--in this case known only to the defendant--key.

Data on a disk is translated by computers but, in its recorded state, it isn't readable in any human sense of the word. Put another way, even people who can read binary can't read the data bits any more than you can hear the grooves on a vinyl record. A machine translates it into something we can understand.

Speaking of translations, it surely wouldn't be tampering with evidence to keep records in the Spanish, Sweedish, or Swahili language? So why not a secret language known only to oneself? And how is that different from encryption?

I don't think PC sec. 37.09 would automatically be broken by a self-destruction mechanism. The statute requires that the actor knows that "an investigation or official proceeding is pending or in progress." If the self destruct mechanism were in place prior to any investigation then it would not seem to trigger the statute.

If you gave the authorities a 'destruct' password after being told they were investigating, that might be tampering. But if they were simply trying to guess the password, and got one that destroyed the evidence, then it I don't think it would qualify. There are destruction designs out there --- type in the wrong password three times and it destroys the drive, etc. They are unusual, however.

Similarly, there are vaults/safes designs that can activate pyrotechnics to destroy contents if an attempt to crack is detected. Once again, not your common thing, but they are out there. A little more common, however, are containers for hard drives that will attempt to degauss (i.e. demagnetize or erase) the hard drive within if unauthorized attempts to open it are detected.

Often, when a drive is seized the techs will physically remove the drive (and even separate the storage media) prior to copying anything. This would get by any software on the computer that would try and erase anything.

Some sort of automatic data self destruct might be difficult to work with. If a defendant triggers it after he knows the police want the data, that is different.

The safe analogy might run afoul of other state or federal laws. Arson maybe?