View article

We met as a group during the Homomorphic Encryption Standardization Workshop on July 13-14, 2017, hosted at Microsoft Research in Redmond. Researchers from around the world represented a number of different communities: government, industry, and academia. There are at least 6 research groups around the world who have made libraries for general-purpose homomorphic encryption available ([SEAL],[HElib],[Palisade],[cuHE],[NFLLib],[HEAAN]) for applications and general-purpose use, and demos were shown of all 6 libraries. All 6 of these general-purpose libraries for homomorphic encryption were based on RLWE-based systems (Ring Learning With Errors), and all libraries implemented one of two encryption schemes (BGV or B/FV) and also displayed common choices for the underlying ring, error distribution, and parameter selection.

Homomorphic Encryption is a breakthrough new technology which can enable private cloud storage and computation solutions. Demos shown at the workshop included a SEAL demo of CryptoNets, which performs efficient computation of image processing tasks such as handwriting recognition on encrypted data using neural nets. Many other applications are described in detail in the white paper by the Applications group. In order for Homomorphic Encryption to be adopted in medical, health, and financial sectors to protect data and patient and consumer privacy, it will have to be standardized, most likely by multiple standardization bodies and government agencies. An important part of standardization is broad agreement on security levels for varying parameter sets. Although extensive research and …