Search

Trusted Computing is a promising approach to improve the security of computer systems. However, current releases of the Java platform do not provide support to utilize the Trusted Platform Module (TPM). This paper presents several building-blocks that lead to the integration of TC into Java. It outlines the issues that arise with multiple TPM-virtualizations in the context of managed environments. Further, it summarizes the design and implementation of Java TC-libraries that support the major operating systems on TPM-enabled platforms, while still considering alternative architectures. The final aspect covered is the ongoing standardization process of a future Trusted Computing API for Java.

The development of trusted systems, as envisaged by the Trusted Computing Group, assumed that the computing environments are uniform in terms of their operational environment, including hardware configuration, execution of a standard set of applications, operating system and facilities and procedures that allow the issue, revocation and maintenance of critical encryption keys and authorization certificates. These assumptions may be applicable to a single managed enterprise infrastructure. However, in situations where the users are mobile, or the computing environment is heterogeneous and the Internet provides the connectivity, the management of trust between enterprises becomes overwhelmingly difficult, if not impossible. As a result, deployment and uptake of trusted secure systems based on Trusted Platform Module have not been as successful as first envisaged. In this paper, we report on our experiences in designing and implementing a prototype personal trusted device called the Trust Extension Device, or TED, that provides users with a portable trustworthy environment for conducting transactions on any Internet connected computer.

Personal Computers (PC) with build in Trusted Computing (TC) technology are already well known and widely distributed. Nearly every new business notebook contains now a Trusted Platform Module (TPM) and could be used with increased trust and security features in daily application and use scenarios. However in real life the number of notebooks and PCs where the TPM is really activated and used is still very small.

Network Access Control (NAC) is the most promising approach to provide protection against sophisticated attacks that first compromise endpoints to subsequently continue their evil work in networks accessible via the compromised endpoint. Trusted Network Connect (TNC) is a NAC approach featuring interoperability and unforgeability due to its openness, broad vendor support and integration of Trusted Computing functions.

Hypervisor technology in combination with tamper-resistant hardware such as Trusted Platform Modules has the potential of providing a significant new level of security for computers. These new systems will be able to protect legacy applications from applications such as malware, and they will enable the creation of new secure applications. This paper analyses the requirements for the graphical user interface (GUI) of such a system. First, application scenarios are presented to illustrate how extended Trusted Computing (TC) systems might be used in the future. Requirements for a trusted GUI are then derived from the analysis of articles on TC found in the media and from a survey of experts in information technology. These requirements are compared to existing proposals, showing how some of them are either inconvenient or insecure in the hands of end users. Based on this analysis, proposals are derived for the design of a trusted GUI suitable for laypersons and administrators, which uses familiar patterns, and yet protects against visual mimicry attacks. These proposals pave the way for hypervisors’ security-enhanced GUI.

Network Access Control (NAC) solutions promise to significantly increase the security level of modern networks. In short, they allow to measure the integrity state of an endpoint that tries to get access to the network. Based upon the measurement results, which are compared to a defined NAC policy, access to the network can be allowed or denied. One problem of all currently available NAC solutions is referred to as the “lying endpoint” problem. Normally, special software components are responsible for gathering the relevant integrity information on the endpoint. If an attacker modifies those software components, an endpoint can lie about its current integrity state. Therefore, endpoints which are not compliant to the defined NAC policy can get access to the network. Those endpoints must be considered as potential threat. This paper summarizes a possible solution for the lying endpoint problem based upon the specifications of the Trusted Computing Group (TCG) and the results of the two research projects TNC@ FHH and Turaya. The goal is to develop an open source, TNC compatible NAC solution with full TPM support within a new research project: tNAC.

Many solutions have been proposed to raise the security level of virtualisation. However, most overlook the security of virtual disk images. With our paper we present a secure, flexible and transparent security architecture for virtual disk images. Virtual machines running on our architecture transparently benefit from confidentiality and integrity assurance. We achieve this by incorporating the concepts of Trusted Computing and in particular the Trusted Platform Module (TPM). This enables us to provide a secure and flexible trusted virtual disk infrastructure to a broad number of platforms. Furthermore, the unique concept of Trusted Virtual Disk Images (TVDI) allows an image owner to stay in control over the disk image throughout its complete life-cycle.