Title: SHORTCOMINGS OF TCP FOR SUPPORTING PRIVACY

Shortcomings of TCP for Supporting Privacy

It appears, however, that with the use of TCP and their application on the service provider’s side a time-related security vulnerability is present, which can be exploited for a violation of the confidentiality of disclosed data. A substantial part of this chapter is the analysis of this vulnerability and its evasion by means of a technical mechanism.

Expressed in simple terms, trusted computer platforms are approaches with which security attributes of a physical component can be extended to a system and its applications. The aim of TC platforms is to guarantee the integrity of the execution of an application and, if necessary, credibly attest this. To achieve this goal, components (also known as trust anchors) are currently used, which cannot be influenced in their operation by software or physical intervention. Assertions and attestations about a platform are based on these. In addition to the logging of application execution, trusted computer platforms mainly ensure active protection of the applications to be executed by realizing confidential and integer memory areas through the application of cryptography.  

TCP applications are thus suitable for giving an outsider credible information about a platform, provided that he is confident that the TCP mechanisms operate as expected. TCP mechanisms can therefore also be regarded as a means for logging and communication and are hence suitable for the transparent execution of remote tasks and responsibilities (see ).

Figure Shifting tasks to remote systems with verification of their orderly execution.

No previously existing trust relationship is thereby required, for example through reputation systems (Cranor and Resnick, 2000; Bohnet and Huck, 2003) between users and providers. This relationship is replaced by the application of additional technology and by trust in the correct mode of operation of this technology.  

It is to be noted, however, that an automatic increase in the security and confidentiality of the applications performed on a TCP does not accompany with their deployment. This is the responsibility of the applications performed and is therefore a software issue. The undecidability theorem attests this among other things. This theorem attests that there are no generally accepted methods which enable the detection of undesired codes in software (Thompson, 2003). The assertion is based on the mappability of the named problem onto the undecidable halting problem. This also applies to Trusted Computing platforms. Assertions about the behaviour of a piece of software are possible, if there has been no impact on the system at any point in time through unknown influences and applications.