TCPA/TCG - Trusted or Treacherous

From ThinkWiki
Revision as of 11:26, 4 July 2005 by Pitsche (Talk | contribs) (minor typo again)
Jump to: navigation, search

Trusted or Treacherous???

You own a ThinkPad build after the year 2000? Or you want to buy a brandnew ThinkPad from Levono (IBM)? Do you want to know, what the Embedded Security Subsystem is doing (or can do) in your ThinkPad?

If you answer one or all of these questions with yes, read on :-) This is about Trusted Computing, TCPA, Palladium, the "Fritz"-Chip, Digital Rights Management and your freedom of choice. At the end of this article you find related hyperlinks to this important topic for computer users.

NOTE!
The following quote is a very short version of the promises and risks of Trusted Computing as the TCG wants it.

After the quote, I will provide you with more informations, but reading only the following quote will give you at least a litle impresion...

"Where's the problem?

It is clear that trusted computing hardware provides security benefits, if software is prepared to take advantage of it. But trusted computing has been received skeptically and remains controversial. Some of the controversy is based on misconceptions, but much of it is appropriate, since trusted computing systems fundamentally alter trust relationships. Legitimate concerns about trusted computing are not limited to one area, such as consumer privacy or copyright issues.
We have at least two serious concerns about trusted computing. First, existing designs are fundamentally flawed because they expose the public to new risks of anti-competitive and anti-consumer behavior. Second, manufacturers of particular "trusted" computers and components may secretly implement them incorrectly.
[...]
Conclusion

We recognize that hardware enhancements might be one way to improve computer security. But treating computer owners as adversaries is not progress in computer security.
The interoperability, competition, owner control, and similar problems inherent in the TCG and NCSCB approach are serious enough that we recommend against adoption of these trusted computing technologies until these problems have been addressed. Fortunately, we believe these problems are not insurmountable, and we look forward to working with the industry to resolve them."

Quote in italic, bold emphasis by me, Source: Electronic Frontier Foundation.

Background:

Recently, the number of known security incidents has been dramatically increasing. Thus, security issues in computer industry have been pushed forward.

The Trustworthy Computing Initiative by Microsoft and other members of the Trusted Computing Group (TCG) are working on a paradigm shift in information technology, which will be the biggest change of the information landscape since decades.

The new concept is to place an especially "trusted" observer (a.k.a. "Fritz"-Chip, a secure cryptographic coprocessor) into information handling devices, to prevent even the device owner from certain operations.

In this context, "trusted" thus means that the owner of the information can trust the device, and verify that the device's "trustworthiness", while on the other hand the device owner no longer has full control over her device.

According of a lot of technical analysis most researchers have fundamental critics on the main design considerations. The new infrastructure will offer only minor protection against worms and viruses. On the other hand Trusted Computing offers a lot of features which can be used to protect the personal computer against the users.

Compared to this, positive features like a more secure hardware storage for cryptographic keys seem to be a very small benefit.

Additionally, the market domination of Microsoft, obscurities regarding the needed trust infrastructure and a heap of patents have lead to critical evaluations from cryptographers, privacy organizations and European institutions.
Because of this pressure the Trusted Computing Group has modifed its proposal. The recent specification is "TCG 1.2".

A short history of TCPA, TCG, Palladium and NSCB:

  • 1999: The Trusted Computing Platform Alliance (TCPA) is founded by Intel, Microsoft, HP, Compaq and IBM.
  • 2002: The TCPA Main Specification Version 1.1b has been published in February 2002.
  • 2003: The Trusted Computing Group (TCG) is founded in April 2003 by AMD, HP, IBM, Intel and Microsoft.
Compared with the TCPA the TCG is less democratic organized and the high membership fees obstruct the possibilities for small companies and non profit organizations to participate.
  • Microsofts own concept for 'Trustworthy Computing', "Palladium", is expected to cost some hundreds of million cash.
  • In the beginning of 2003 the name was changed to "next-generation secure computing base" (NGSCB).
  • In summer 2004 the NGSCB effort seems to have stopped.

What is all the name changing about?
Some say, that the change was a reaction to the negative publicity, because Palladium and TCPA was soon equated with a 'Nineteen Eighty-Four'-Scenario of an 'Orwellian society', the ever-present, all-seeing 'Big Brother' and other privacy issues.

TCG Hardware Architecture

The "Trusted Platform Module (TPM)" (a.k.a. "Fritz"-Chip) is the central building block of the TCG architecture and the first implementation can be seen as just a hardwired smart card.

There are also discussions to integrate the whole functionality into the main processor, which would increase resistance against tampering attacks (see also Intel "LaGrande").

The most important services of the TCG specifications are:

  • Hardware storage for cryptographic keys
  • Secure booting
  • (Remote) Platform Attestation (meaning that somebody can check the state of your personal computer over the internet)
  • Sealing (meaning binding data to a specific platform and application)

Generally: There are good arguments that these features can be used to improve the security of computer systems.
But: Some of these features can already be established by todays smart card supported systems!

Drawback 1: Remote Attestation is a good feature to remotely detect tampering of the computer, as long as this 'somebody' is the owner of the platform.
But: If this Remote Attestation is used by third parties, serious privacy and market domination issues arise.

Drawback 2: There are certainly legitimate reasons for Sealing.
But: The main use case seems to be consumer-unfriendly new 'business cases' for content dealers which involve locking down content to a single platform, based on connecting content to a specific device without any migration options. This means e.g. if the user wants to use his music to a portable player he should be forced to buy another license. It seems to be doubtful, if customers will enjoy this limitations.

In any case the possible problems of giving away control of the personal hardware should be evaluated carefully.

There have also been two important critiques regarding the hardware security of the "Trusted Platform Modules (TPM)".

The first one has been the insuficient security certification against hardware attacks. In TCG 1.2 this critique has been addressed by an improvement of the hardware requirements and it has to be seen how strong the resistance against sophisticated attacks at intensively daily usage will be.

The second one addreses the 'black box'-characterisitcs and therefore 'hidden channels' in the TCG-Hardware. Hidden channels smuggle secret information to third parties and it has been a well known fact for many years, that hidden channels are easy to implement in black box hardware.

TCG and Digital 'Restrictions' Management

As stated above, the philosophy behind Remote Platform Attestation and Sealing seems to be a protection of the computer system or electronic device against its user and owner.

What will do this to the use of digital media content on electronic devices? The answer of the IT- and the Entertainment-Industry is "Digital Rights Management" or just shortly "DRM".

The DRM component takes control over the rest of the user's device which they rightfully own (e.g. MP3-Player or a ThinkPad) and restricts how it may act, regardless of the user's wishes (e.g. preventing the user from copying a song). All forms of DRM depend on the device imposing restrictions that cannot be legally disabled or modified by the user. In other words, the user has no choice.

So a new 'name' for DRM came up: Digital Restrictions Management instead of Digital Rights Management.

I cannot go too much into the details here, so please have a look at the article in the Wikipedia Encyclopedia, where I took the quote from. The bold emphasised addition is mine.

Censorship and Avoiding Whistle Blowers

The Siamese twin of Digital Restriction Management is censorship. The same techniques which avoid copying music songs can be used to limit the access to all kinds of documents. The combination from DRM and observation hardware (TCG is in your mind?) leads to very dangerous implications.

Giving a real world example, the Chinese government could easily block the use of all documents containing the words "Dalai Lama" on 'trusted' computer systems.

Another application is the fight against whistle blowers. E.g. government documents about the deportation of own citizens to countries with a doubtful law system or about supporting illegal wars could made only readable for government computers and combined with a expiration date. This might make it very dificult for the society or following generations to disclose these breaches of humanity.

Open Source Software and TCG

Since Microsoft controls a overwhelming part of the OS market, it seems to be rather dificult to evaluate the TCG proposal separated from the Palladium project.

TCG versus GPL: At least two companies are researching on "TCG-enhanced" versions of GNU/Linux. According most security researchers it seems to be necessary to evaluate programs which have access to the 'trusted part'.

This has huge implication for the development of free software. Following a possibly expensive evaluation there will be a signature for one program version.

Even if the program stays under GPL every change of the code will make the signature invalid. This seems to be a strong violation of the main philosophy of Open Source software.

Embedded Security System (1.0)

Embedded Security System (in IBM documents there is no use of the additive version-nummer 1.0) is using the heayvily disputed "TCG 1.1"-specification.

(Features etc. will soon be added here - your help, support or cooperation is very much appreciated)

Embedded Security System 2.0

The recent TCG-specification is "TCG 1.2" and Embedded Security System 2.0 is supposed to use this newer specification.

(Features etc. will soon be added here - your help, support or cooperation is very much appreciated)

Summary

There are still a lot of critical questions, even though TCG 1.2 contains many steps into the right direction.


Related Links


Read more at Wikipedia, the free encyclopedia:


ThinkPad-Models with TCPA:

TCPA was introduced as the so called "Embedded Security System" for the first time in the ThinkPad T23 (July 2001).

Almost every ThinkPad, which was build after the T23-Series is equipped with "Embedded Security System" or "Embedded Security System v2.0" - except for the A30-Model (but the A30p has a "Embedded Security System"!!!).

If you want to buy a renewed or used ThinkPad with a Pentium III-CPU, but don't want TCPA inside, than you can choose between the A30 or every model version of the A22-, A21-, A20-, T21-, T20-Series or the 240X, the 570E or the 600X (errors may occur, so please check for yourself before buying!).