Slashdot Mirror
Army to Require Trusted Platform Module in PCs
Overtone writes "Federal Computer Week is reporting that the U.S. Army will require hardware-based security via the Trusted Platform Module standard in all new PCs. They are a large enough volume buyer that this might kick start an adoption loop."
Army requires TMP so that it can circumvent single-vendor prohibition and be Intel(R) only.
The question still remains whether the user himself can trust the trusted computing platform.
If your government or seller or whatever doesn't trust you, doesn't even try in the least, how the hell are you supposed to trust him? The most logical path would be to fully distrust him. And therefore to distrust and refuse trusted computing platform.
"The way we can tell it's C# instead of Haskell is because it's nine lines instead of two." -- wadler
If I am hanging from a rope over a cliff, I Trust the rope. I "Entrust it with my security" whether or not I find it worthy of that trust.
I cried real tears when Li Mu Bai died.
I personally abhor the notion of Trusted Computing on my personal computer, but if you're using a computer provided to you by the government or a corporation for the express purpose of working, it's their right to control what goes on on that computer. It's possible that this will help to stem the tide of malware (at least in corporate environments) by rejecting execution privledges, and allow IT staff to better enforce policies about what can and cannot be run on their computer. It would also help stop things like the Free USB Key Attack (formerly discussed on slashdot).
Of course, this could also make users feel like they are not trusted, and could even lead to overconfidence in the security of the system. Still I see it as a major plus, at least unless I get saddled with it at home.
This is a worrying scenario. Apart from the minor issue that external users will not want to pay for the dongles and that the internal customer is seeing his IT bill spiral, Trusted Computing seems to be heading to a Mexican standoff situation as follows:
Device 1: Permit me to inspect your system by downloading and running this program.
Device 2: Only after YOU have allowed me to verify your credentials by uploading and running this program.
Device 1: No, it is I who am deciding whether you are to be trusted!
Device 2: No, it is I who am deciding that!
Device 1: Anyway, my content is digitally signed by Microsoft, and you must trust it.
Device 2: Microsoft? Not a hope in Hell. I require all downloads to be digitally signed by Steve Jobs in person with a DNA signature.
And so on. Quis custodiet ipsos custodes? And how long before an army unit gets wiped out because of a defective dongle?
Pining for the fjords
It makes sense for the Army to require TCP. Stolen/lost laptops wouldn't immediately result in a security leak. But this can be achived cheaper, quicker and (and here comes the key point) with more control on the Army's side. Linux can encrypt documents just the same way TCP wants to offer, the difference lies in the open source concept: This inherently gives you the ability to check the security (provided you can read code, but I guess the Army can afford hiring someone who does) of your system.
TCP requires you to trust the person/group that made the security for you. You put yourself completely into the hands of the corporation(s) that create your TCP platform, and you are fully dependent on their ability to come up with a good protection scheme. Not to mention that you have to trust them, implicitly, that they do not want to spy on you and that they are better than their adversaries.
With TCP you hand over the responsibility for security. But you also hand over control. And it has the potential to lure you in a false sense of security which invariably leads to slacking. More than once I've seen a behaviour of neglect in a high security area (I've had my share of time in that field), with people relying so heavily on the technical implementations that they forgo the most basic security measures called for by common sense, because "Hell, what DO we have that security concept for, if I can't trust it fully?"
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
A country's armed forces ought to have the power to demand the full source code of every application running on their computers, and the resources to write all their own software wherever necessary. There is no shortage of Open Source applications they could use for starting points .....
Je fume. Tu fumes. Nous fûmes!
That's a total lie. Almost everything in that piece of propaganda masquerading as a FAQ is a lie.
If you want the truth about TC, try Seth Schoen of the EFF. He has a good summary in his recent blog entry:
OK, I'll give a car analogy. They suck, but are fun. My '85 Buick Elektra (I still miss him) was a Trusted Transportation Platform.
Well, I think a correct car analogy for Trusted Computing would be not YOUR car but your DADDY's car. You would trust your daddy to issue you the keys when you needed and your daddy would trust you not to damage the vehicle. Of course, any time there would be any conflict between you two ("dad, I swear to God that this scratch was here before!"), daddy would have the ultimate saying ("swear to anyone you want, kid, but you're gronded").
And you could only trust your dad won't abuse his power. TPM is the same provided that you trust Microsoft, Apple et al love you like your parents.
You're quite right of course. If the "resistance" in Iraq confined its attacks to America soldiers, they would be freedom fighters. In reality, attacks on American troops are rare. They mostly target other Iraqis who simply aren't the "right" type of Muslim. That barely even qualifies as terrorism; it's more along the lines of a slow, decentralized holocaust.
Imagine if the French resistance in WW2 had schismed into seperate Catholic and Protestant factions, and they'd spent all their time killing each other instead of collecting useful intelligence for the Allies. The people of Yugoslavia put aside enormous cultural difference, ceased all internal violence, and totally unified to form the largest and strongest resistance army that there has even been -- and ousted the Nazis themselves. Tito and company -- probably the best example of freedom fighters since the American war of independence. By way of contrast, consider China during WW2. If the Chinese had cooperated, Japan would have never been able to successfully invade let alone retain control once they were in. Chinese resistance failed because imperialists and Maoists were never able to put their own civil war on hold (although the Maoists apparently tried several times, which part of the reason that the people supported them after the war). It is just mind boggling how far the Iraqi extremists are from being anything other than a plague upon their homeland.
This would be a really worrying thing, but the fact is TPM has already won. It won the instant that Apple adopted TPM and the communities who were publicly worrying and complaining about Palladium and Trusted Computing for all those years went suddenly silent and shrugged the instant that nebulous notions like "freedom" came into conflict with solid, purdy white plastic.
Here is the thing: TPM's adoption was waiting not on an adoption cycle exactly, but an apathy cycle. TPM was never something that the consumer was supposed to approve of, want, or even really know was there. The adoption of TPM was mostly counting on the consumer not having any idea what they were buying, counting on the blinking 12:00 effect, counting on the idea that most consumers would not even know TPM was in their computer until the first time that they try to do something and the computer says "no".
TPM isn't there for the consumer. It's there to protect the computer from the consumers. It's there to allow software and content vendors to trust your computer, to trust your computer to ensure it will act in their interests and not yours. These vendors are the ones that TPM is being done for the benefit of, not the consumer. This means that in order for TPM to win, it isn't necessary for the consumer to "adopt" it. All that has to happen is for the consumer to fail to actively reject it when it is quietly dropped into the hardware they were going to buy anyway.
And that's already happening. So although the military would legitmately represent an adoption cycle-- the military, of course, has a legitimate and logical need to create networks within which the machinery is trusted and the user is absolutely not-- it doesn't really matter. The military isn't the kind of adoption TPM needs to reach enough critical mass that vendors can begin requiring it in new applications, I don't think-- it's not like military hardware is going to be used to run lots of games and DRMed consumer media, as far as I know. The worrying thing is TPM's level adoption in the consumer segment, since that's where it has potential to do actual harm. And that's already begun, and so far nothing is happening to stop it...
Irritable, left-wing and possibly humorous bumper stickers and t-shirts