The knowledge in technologies needed to build secure platforms, or Security Architectures, has significantly matured over the recent years. These include small interface technologies, access-control contracts, tunneling, secure booting, effective resource control, and virtual machines. Putting together these ingredients into a small secure platform seems straightforward, yet still remains to be done, and has the potential of making operating systems more dependable.

Many applications of cryptographic identification protocols are vulnerable against physical adversaries who perform real time attacks. For instance, when identifying a physical object like an automated teller machine, common identification schemes can be bypassed by faithfully relaying all messages between the communicating participants. This attack is known as mafia fraud.

Microsoft Palladium (Pd) and TCPA are announced to be the next-generation computing platforms, and claimed to improve users' security. However, the public debate on TCPA/Pd is full of skepticism and mistrust about the claimed security enhancements for the users. People are concerned about those features and capabilities of these systems that can be applied to realize Digital Rights Management: they may allow content providers to gain too much power and control over the use of digital content and users' private information. The general negative and sometimes prejudiced response about TCPA/Pd left less space for an unbiased and objective evaluation.

Over the last two years, the computing industry has started various initiatives to increase computer security be means of new hardware. The most notable effort is the Trusted Computing Group (TCG, former TCPA), which unites most important companies to develop a standard for security hardware. While efforts towards more security are generally appreciated, the new technology has raised fears that the user will lose control over his platform, and that its extensive use would harm competition in several areas of IT technology. In this paper, technical modifications to the TCPA Specification (based on version 1.1b) are suggested. If implemented, these suggestions would eliminate the potential to abuse the technology against the users. To the best of our knowledge, none of the intended use cases of the TCG would be affected. The proposals will be discussed with TCG Members on the GI annual conference 2003 in Frankfurt and then officially presented to the TCG.