Trusted Computing aims to reduce the vulnerability of personal computers to viruses and malware by incorporating tamper-resistant processors. Among others, such processors can be used to protect cryptographic keys, to securely boot operating systems, and to prevent programs from reading or writing each other’s memory. Market analysts expect that by 2010 the majority of personal computers for sale will include Trusted Computing chips that implement the specifications of the Trusted Computing Group (TCG).

While Trusted Computing undeniably has the potential to greatly improve the security of individuals and organizations, it is highly controversial. There are good reasons for this: the current TCG specifications raise severe concerns relating to privacy, autonomy, and civil liberties. While the TCG has taken some engineering steps in an attempt to address some of them, the current specifications still fall short.

The U-Prove™ technology addresses many of the concerns while at the same time expanding the benefits of Trusted Computing for both users and organizations:

• Private remote attestation: Remote attestation (an important TCG feature) allows a remote party to verify that it is communicating with a device that contains a valid Trusted Computing chip. The TCG specifications describe two methods to electronically verify the presence of a Trusted Computing chip without disclosing tracing information to the verifier. One TCG approach relies on a trusted third party (called the Privacy Certification Authority). The other approach does away with this trust requirement by having Trusted Computing chips perform a complex zero-knowledge cryptographic proof (this is referred to as direct anonymous attestation). Unfortunately, neither approach really protects privacy: direct anonymous attestation requires users to trust that their Trusted Computing chips generate random numbers that are not known to their manufacturer or designated third parties, a claim that cannot be verified. The U-Prove technology, in contrast, enables remote attestation without computer owners having to trust the proper functioning of their Trusted Computing chips in any way, even in the face of active collusions between chip manufacturers and attestation verifiers; individuals need merely trust the software that is run by their own CPUs (which may be freely obtained and can be inspected by anyone).
• Efficient binding of identity assertions: An unlimited supply of protected identity assertions can be electronically bound over open networks to a previously issued Trusted Computing chip. Trusted Computing chips can enforce any third-party security policies (whether of their own users, of assertion issuers, of verifiers, and/or of other parties) throughout the lifecycle of the identity assertions. At the same time the chips cannot leak any unwanted information to the outside world and cannot learn any of the information contained in the identity assertions they help protect (unless it is expressly provided to them by their own users). A highly constrained chip (e.g., an 8-bit chip) suffices to protect billions of identity assertions; no expensive cryptographic operations need be performed by Trusted Computing chips. This performance feature also helps to minimize the risk of side-channel attacks (such as electromagnetic analysis and differential power analysis).