PRISMACLOUD aims at bringing novel cryptographic concepts and methods to practical application to improve the security and privacy of cloud based services and make them usable for providers and users.
The purpose of this report is to document the progress on research activities within the Task 4.3 Privacy enhancing cryptography in the second year of the PRISMACLOUD project. We thereby focus on privacy-preservation for users of cloud services as well as service providers. In particular, we will improve and propose privacy-enhancing cryptography such as signature schemes for constructing anonymous credentials as well as group signature schemes for the cloud environment with a focus on user's access privacy in authentication and authorization, private billing for the use of cloud services as well as privacy for cloud providers enabling them to selectively prove properties about their certiFIed infrastructure without disclosing the blueprint of their infrastructure.
To this end, this task conducts research in the following fields.
4.3.1 Privacy-Preserving Cryptography for the Cloud. In this task, we will investigate privacy-preserving cryptographic protocols and in particular anonymous credential systems and group signature schemes. Most such privacy preserving schemes as (updatable/stateful) anonymous one-show/multi-show credentials, or group signatures are obtained by means of (generic) transformations from signature schemes enjoying specific properties (such as blind/partially blind signing support, support for signing commitments, randomizability and compatibility with efficient zeroknowledge proofs). We will on the one hand perform research in anonymous credential systems that do not follow the traditional proof-of knowledge paradigm, but are based on alternative constructions (such as ideas from malleable signatures), which make them conceptually simpler as well as to integrate additional features such as a state and updateability. Furthermore, we will investigate these approaches focusing on identifying difficulties and trade-offs that have to be made when targeting for implementations in resource constrained hardware. In this deliverable we present three publications related to this task.
4.3.2 Certified and Verifiable Infrastructure for Cloud Services. In this task we develop and optimize for practical use in virtualized infrastructures is a signature scheme on committed graphs with a zero-knowledge proof system. Such a scheme allows an auditor to analyse the configuration of a cloud, and issue a signature on its topology. The signature encodes the topology as a graph in a special way, such that
the cloud provider can use it to prove in zero-knowledge high-level security properties such as isolation of tenants to verifiers, such as the tenants, without disclosure of secret information. By that the verifying tenant can be confident that the infrastructure is congured securely as promised by the provider and be assured at the same time that no information about his resource pool is leaked to other tenants. In this deliverable, we present research on proactive security analysis of virtualized infrastructures based on a graph representation aligned with the topology certication pursued in PRISMACLOUD. We present research to establish minimal functional units with tightly controlled capabilities that can then by certified as vertices in the topology certification. We further describe research on the certification of geo-location for physical systems in virtualized infrastructures and the on proofs of geo-location separation in zero-knowledge. We report on research on the representation of directed graphs for the topology certication of advanced security properties and the expressivity of such schemes.