Cingulata is used to develop applications capable of processing encrypted data without having to decrypt it. With Cingulata, processing data in the cloud or in collaborative workflows no longer has to compromise data privacy.
Cingulata is a compilation chain that can process encrypted data without having to decrypt it. It is useful in collaborative workflows, for example, for applications requiring the processing of customer or partner data without compromising the privacy of the information the data contains.
Using fully homomorphic encryption, an advanced encryption technique, Cingulata creates, optimizes, and deploys applications capable of processing encrypted data on remote servers, guaranteeing end-to-end privacy. With Cingulata, partners can exchange and process confidential data without revealing the information the data contains. It can also be used to develop cloud-based services that can analyze encrypted confidential data, without revealing the information in the data or the results of the analysis.
This open-source software meets a wide range of needs in the defense, energy, health, biometrics, multi-mode transportation, manufacturing, and other industries.
Cingulata offers several major advantages:
Scalnyx partnered with CEA-List to develop a secure-by-design solution to speed up processing on encrypted data. Performance is still a major challenge for this type of processing. Scaltrust, the solution developed by Scalnyx, leverages CEA-List’s Cingulata compilation chain and Scalnyx’s software acceleration technology.
Learn more about the R&D partnership between Scalnyx and the CEA.
Revisiting stream-cipher-based homomorphic transciphering in the TFHE era, A. A. Bend oukha, A. Boudguiga and R. Sirdey. Foundations of Privacy & Security (FPS), 2021. « in press ».
SPEED: Secure, PrivatE, and Efficient Deep learning, A. Grivet Sébert, R. Pinot, M. Zuber, C. Gouy-Pailler and R. Sirdey. Machine Learning Journal 110, 675–694, 2021 (DOI). Presented at ECML’21.
Efficient homomorphic evaluation of k-NN classifiers, M. Zuber and R. Sirdey. Proceedings on Privacy Enhancing Technology Symposium 2, 111-129, 2021 (DOI). Presented at PETS’21.
Faster homomorphic encryption is not enough: improved heuristic for multiplicative depth minimization of boolean circuits, P. Aubry, S. Carpov and R. Sirdey. 2020 Cryptographers’ Track at the RSA Conference.
Automatize parameter tuning in Ring-Learning-With-Errors-based leveled homomorphic cryptosystem implementations, V. Herbet. 2019.
Practical fully homomorphic encryption for fully masked neural networks, M. Izabachène, R. Sirdey and M. Zuber. Proceedings of the 18th International Conference on Cryptology and Network Security, 24-36, LNCS 11829, 2019 (DOI).
A SaaS implementation of a new generic crypto-classifier service for secure energy efficiency in Smart Cities, O. Stan, M.-H. Zayani, R. Sirdey, A. Ben Hamida, M. Mziou-Sellami and A. Ferreira Leite. In Smart Cities, Green Technologies and Intelligent Transport Systems, pp. 90-115, 2019 (DOI).
New Techniques for Multi-value Input Homomorphic Evaluation and Applications, S. Carpov, M. Izabachène, V. Mollimard. 2019 Cryptographers’ Track at the RSA Conference (DOI).
Go to Cingulata
