User-centric identity as a service-architecture for eIDs with selective attribute disclosure

Unique identification and secure authentication of users are essential processes in numerous security-critical areas such as e-Government, e-Banking, or e-Business. Therefore, many countries (particularly in Europe) have implemented national eID solutions within the past years. Such implementations are typically based on smart cards holding some certified collection of citizen attributes and hence follow a client-side and user-centric approach. However, most of the implementations only support all-or-nothing disclosure of citizen attributes and thus do not allow privacy-friendly selective disclosure of attributes. Consequently, the complete identity of the citizen (all attributes) are always revealed to identity providers and/or service providers, respectively. In this paper, we propose a novel user-centric identification and authentication model for eIDs, which supports selective attribute disclosure but only requires minimal changes in the existing eID architecture. In addition, our approach allows service providers to keep their infrastructure nearly untouched. Latter is often an inhibitor for the use of privacy-preserving cryptography like anonymous credentials in such architectures. Furthermore, our model can easily be deployed in the public cloud as we do not require full trust in identity providers. This fully features the Identity as a Service-paradigm while at the same time preserves citizens' privacy. We demonstrate the applicability of our model by adopting to the Austrian eID system to our approach.