Desktop version

Home arrow Computer Science arrow A Practical Guide to TPM 2.0

Quick Tutorial on TPM 2.0

This chapter describes the major uses of TPM capabilities. The use cases for which TPM 1.2 was designed still pertain to TPM 2.0, so we begin by exploring those use cases and the designed functionality that enables them. Then we move to new aspects of the TPM 2.1 design and the use cases enabled by those capabilities.

As noted in Chapter 1, the rise of the Internet and the corresponding increase in security problems, particularly in the area of e-business, were the main driving forces for designing TPMs. A hardware-based standardized security solution became imperative.

At the same time, due to the lack of a legacy solution, security researchers were presented with a golden opportunity to design a new security system from the ground up. It has long been a dream of security architects to not merely patch problems that existed in earlier designs, but also provide a security anchor on which new architectures can be built.

The TPM 1.2 specification was the Trusted Computing Group's (TPG's) first attempt to solve this problem and was aimed at addressing the following major issues in the industry:

Identification of devices: Prior to the release of the TPM specification, devices were mostly identified by MAC addresses or IP addresses—not security identifiers.

Secure generation of keys: Having a hardware random-number generator is a big advantage when creating keys. A number of security solutions have been broken due to poor key generation.

Secure storage of keys: Keeping good keys secure, particularly from software attacks, is a big advantage that the TPM design brings to a device.

NVRAM storage: When an IT organization acquires a new device, it often wipes the hard disk and rewrites the disk with the organization's standard load. Having NVRAM allows a TPM to maintain a certificate store.

Device health attestation: Prior to systems having TPMs, IT organizations used software to attest to system health. But if a system was compromised, it might report it was healthy, even when it wasn't.

The TPM 2.0 implementations enable the same features as 1.2, plus several more:

Algorithm agility: Algorithms can be changed without revisiting the specification, should they prove to be cryptographically weaker than expected.

Enhanced authorization: This new capability unifies the way all entities in a TPM are authorized, while extending the TPM's

ability to enable authorization policies that allow for multifactor and multiuser authentication. Additional management functions are also included.

Quick key loading: Loading keys into a TPM used to take a relatively long time. They now can be loaded quickly, using symmetric rather than asymmetric encryption.

Non-brittle PCRs: In the past, locking keys to device states caused management problems. Often, when a device state had to go through an authorized state change, keys had to be changed as well. This is no longer the case.

Flexible management: Different kinds of authorization can be separated, allowing for much more flexible management of TPM resources.

Identifying resources by name: Indirect references in the TPM 1.2 design led to security challenges. Those have been fixed by using cryptographically secure names for all TPM resources.

TPM 1.2 was a success, as indicated by the fact that more than 1 billion TPMs using the 1.2 specification have been deployed in computer systems. TPM 2.0 expands on TPM 1.2's legacy. Currently, many vendors are developing implementations for TPM 2.0, and some are shipping them. Microsoft has a TPM 2.0 simulator that can also act as a software implementation of TPM 2.0. Some vendors are in the process of sampling hardware TPMs, and other companies are working on firmware TPMs.

< Prev   CONTENTS   Next >

Related topics