Slashdot Mirror
Dartmouth Project Combines Linux With TCPA
SiliconEntity writes "A new project from Dartmouth College demonstrates significant advances in combining Linux with TCPA. The software turns a Linux PC into a 'virtual secure coprocessor', which is able to check that none of its software is compromised and even (in a future version) prove its integrity to a remote system. Full GPL source code is available for the 2.4 kernel.
This work is separate from the earlier IBM research which also combined Linux with TCPA, with the new project apparently more complete and with a road map towards a very functional Linux based trusted computing system. This could be an important technology for Linux to challenge Microsoft as it pushes forward with NGSCB (aka Palladium)."
The TCPA is a comitee and is not something that belongs to Microsoft, although they are part of this comitee. IBM are also working on a TCPA technology. Palladium, or whatever it is called now, is perhaps the most "famous", but definately not the only one.
We want to fight Palladium by fighting acceptance of the idea that the computer should control the user and how he can access the data on his own machine, NOT by developing something functionally equivalent that happens to run under Linux.
Building a DRM system of our own, even if it is open and standards based, just strengthens the paradigm that will leed to an Internet where no data can be accessed as plaintext, applications that are allowed read data have to be accepted and certified by the media industry, and computers exist no longer to enable, but to control, their users.
Please protest against Palladium, TCPA, and all the other DRM proposals by refusing to have anything to do with them: not by strengthening their hand.
(And before somebody replies that TCPA isn't about DRM: Bullshit! Look up what an "endorsement key" is in the TCPA vocabulary.)
Not true at all. DRM and other user control systems only need to be closed when they are software based, because otherwise people can change the programs to remove the user hostile code.
The difference between Palladium and TCPA is really that while Palladium is a whole system for a building user hostile computers, TCPA is just an enabler.
What TCPA does is sign a hash of the OS that is loaded with an "endorsement key", embedded in the TCPA by the vendor and unaccessible to the user. Thus the TCPA chip is a able to do two things: it can verify to an outside source (that trusts the vendor) that the machine is a running a specific operating system (ie one that supports DRM and thus can be "trusted"), and it can encrypt data from one operating system so that another operating system cannot decrypt it.
TCPA provides everything that is needed at the hardware level to write any user hostile system on top of it, because the successive verification of signatures prevents any tampering with the code (even if the OS is open sourced). Palladium could be implemented with TCPA as it's only hardware aspect.
Thus, the argument that is sometimes seen here that TCPA would prevent the computer from booting Linux or any other operating system is false (incorrect scare tactics against these systems are unfortunate, they do more harm then good). What TCPA will do, is enable sites on the Internet to not allow you to read the data they give out, unless you are running an operating system that is user hostile and DRM friendly (and not in the "this site doesn't support mozilla" fashion, which can always be hacked around, but in a cryptologically safe fashion).
It's not meant for you, none of this technology has anything to do with _your_ security. These products are intended to protect people from you, specifically, in this case, the movie industry who don't want you re-recording movies from the monitor cable.
There are two reasons for wanting this in hardware, as opposed to just in the software:
The second reason is a tiny capstone on a pyramid of security that most people haven't built to anywhere near the height where it would be useful. It can be practically disregarded.
All the other things you list can be done without hardware support, and the only catch is that the end user can choose to disable them. Even then, he might need to open up the box to do it. (password-protected BIOS, no booting except from hard disc: most PCs can do that.)
And you're wrong about worms. In most cases, as far as the OS is concerned, the worm isn't running. Some ordinary program (e.g. SQL Server in the case of the slammer worm) is running, but the worm, by feeding it bad data, has caused it to corrupt itself so that it has effectively become the worm. There is no "worm.exe" for a security processor to refuse to run.
But if you don't want to run a particular program (such as a worm), don't run it. There is no need for all this signature stuff, except to prevent the user from running software of his own choosing.
If you did decide to run only code signed by a trusted key, the only reasonable system would be for the owner of the PC to posess that key. (This could be the company IT department, or the individual user for home systems.)
-- Ed Avis ed@membled.com
True .. but tell me:
1) Of what use is a Linux system, if no content can be decrypted on it?
Not much.
2) Will content-providers make content available to versions of Linux which can't be "trusted"?
Undoubtably not. But what format they release the data in is their concern.
It is important to remember that the only political issue here is fighting laws against compulsary DRM and laws against circumventing it where it exists. We should not fall into the whiner trap of trying to claim that we are somehow entitled to "content" in open formats. We are not.
The manner in which we should fight DRM is to explain to be people why they should not accept it. (And we need to start here on Slashdot - look at how many Slashdotters laud iTunes).
3) If you make a "trusted" version of Linux, will it then be modifiable by the user (say, a new kernel-patch)?
It will be modifiable of course, but then you are back to 1).
4) Of what use are Open Source advantages, if you cannot use them?
Not much.
5) Is this a threat to the Open Source development model?
Definitely.
Yes, but you need a root key that is signed by some authority (the kind of keys that are embedded in the chips).
If you can get ahold of one of these keys, then you can simulate running a "trusted" system and cheat the DRM. They won't be easy to get ahold of though. Modchips will probably prove a better avenue.
The TPM is a hardware component that implements the security model. It so happens that this exists on a bunch of modern IBM laptops. It is disabled by default.
Background: The TPM contains a number of PCRs. These are (roughly) hashes of bits of code -- the BIOS, the bootloader, the kernel, etc. The TPM also contains a private/public key pair which is generated when you reinitialize the TPM (i.e. the private key is not known to anybody).
The TPM can be used to encrypt a blob of data using the private key. It can also mark the encrypted blob such that it will only decrypt it if (some set of) the PCRs have the *same* value.
What is this good for?
This means that you can tell if your kernel has been modified in a very secure way. If your application is stored encrypted on disk, then you can ask the TPM to decrypt it (probably you just ask it for the key). It will only perform this operation *if* the boot process was the same as when the application was setup.
It means that someone with a boot floppy cannot get to your data (different boot process). You could also arrange to have the data protected from single-user mode.
However, there is a downside -- upgrading the OS becomes really tricky!
Visited the NYT lately? How about LA Times? How about MIT Press? There are already hundreds, if not thousands of sites, locking their content away behind logins - they don't need DRM to do it.
You're avoiding the point. They already use logins today, and will in the future. But someday they can have these logins protected by DRM technology. They will get a minor economic advantage from this extra protection, but newspaper margins are slim, so they'll grab for it.
Then, it will be impossible to visit those sites with an untrusted OS. It will be impossible to build a PC, compile Linux, compile Mozilla, and use that to browse the web. The freedom of disorganized amateurs to create useful computer systems will be gone.
When free expression is no longer possible on US soil, US dollars will make sure there's a world of domains out there where speech remains free
That's a head-in-the-sand argument. "The government cannot now enforce a prohibition against a behavior. Therefore they will never be able to prohibit it."
Sorry, but in the face of ever-increasing computer power, that viewpoint just doesn't hold up. If you don't believe me, Lessig has published extensive documents describing exactly why.