Abdelrahaman Aly

In this work, we introduce FANNG-MPC, a versatile secure multi-party computation framework capable to offer active security for privacy-preserving machine learning as a service (MLaaS). Derived from the now deprecated SCALE-MAMBA, FANNG is a data-oriented fork, featuring novel set of libraries and instructions for realizing private neural networks, effectively reviving the popular framework. To the best of our knowledge, FANNG is the first MPC framework to offer actively secure MLaaS in the dishonest majority setting.FANNG goes beyond SCALE-MAMBA by decoupling offline and online phases and materializing the dealer model in software, enabling a separate set of entities to produce offline material. The framework incorporates database support, a new instruction set for pre-processed material, including garbled circuits and convolutional and matrix multiplication triples. FANNG also implements novel private comparison protocols and an optimized library supporting Neural Network functionality. All our theoretical claims are substantiated by an extensive evaluation using an open-sourced implementation, including the private inference of popular neural networks like LeNet and VGG16.

The increasing concern for identity confidentiality in the Smart City scenario has fostered research on privacy-preserving authentication based on pseudonymization. Pseudonym systems enable citizens to generate pseudo-identities and establish unlinkable anonymous accounts in cloud service providers. The citizen's identity is concealed, and his/her different anonymous accounts cannot be linked to each other. Unfortunately, current pseudonym systems require a trusted certification authority (CA) to issue the cryptographic components (e.g. credentials, secret keys, or pseudonyms) to citizens. This CA, generally a Smart City governmental entity, has the capability to grant or revoke privacy rights at will, hence posing a serious threat in case of corruption. Additionally, if the pseudonym system enables de-anonymization of misusers, a corrupted CA can jeopardize the citizens' privacy. This paper presents a novel approach to construct a pseudonym system without a trusted issuer. The CA is emulated by a set of Smart City service providers by means of secure multi-party computation (MPC), which circumvents the requirement of assuming an honest CA. The paper provides a full description of the system, which integrates an MPC protocol and a pseudonym-based signature scheme. The system has been implemented and tested.

Business application owners want to outsource data storage, including sensitive data, to the public cloud for economical reasons. This is often challenging since these businesses are and remain responsible for regulatory compliance and data protection, even though cloud providers may do their best to offer (data) protection. Meanwhile, data protection techniques evolve and get better because of continuous research and improvement of advanced encryption. Numerous cryptographic tactics have been proposed, e.g., searchable symmetric encryption (SSE) and homomorphic encryption (HE), that support search and aggregation functions on encrypted data. Each of these tactics has a trade-off between security, performance and functionality, but there is no one-size-fits-all solution. For the application developer, the underpinning concepts of these tactics are complex to comprehend, complex to integrate in a distributed application, and prone to implementation mistakes. In this paper we present DataBlinder, a distributed data access middleware that provides crypto agility by means of configurable fine-grained data protection at the application level. DataBlinder supports adaptive runtime selection of data protection tactics, and offers a plugin architecture for such tactics based on a key abstraction model for protection level, performance and supported query functionality. We have developed this middleware in close collaboration with businesses that face these challenges and offer cloud-based applications in e-finance, and e-health, by implementing and integrating state-of-the-art cryptographic schemes to Dat-aBlinder. This paper illustrates the case of medical data protection with FHIR-compliant [30] medical data. CCS Concepts • Security and privacy → Management and querying of encrypted data; • Information systems → Middleware for databases.

Smart grid allows fine-grained smart metering data collection which can improve the efficiency and reliability of the grid. Unfortunately, this vast collection of data also impose risks to users’ privacy. In this paper, we propose a novel protocol that allows suppliers and grid operators to collect users’ aggregate metering data in a secure and privacy-preserving manner. We use secure multiparty computation to ensure privacy protection. In addition, we propose three different data aggregation algorithms that offer different balances between privacy-protection and performance. Our protocol is designed for a realistic scenario in which the data need to be sent to different parties, such as grid operators and suppliers. Furthermore, it facilitates an accurate calculation of transmission, distribution and grid balancing fees in a privacy-preserving manner. We also present a security analysis and a performance evaluation of our protocol based on existing multiparty computation algorithms.