Frequently Asked Questions

Fully Homomorphic Encryption (FHE) is a type of encryption that allows computation on encrypted data without needing to decrypt it first. The results of these computations remain encrypted and can only be decrypted with the correct key. This means that sensitive data can be processed securely, preserving privacy.

Craig Gentry proposed the first FHE scheme in 2009 as part of his Ph.D. thesis. His approach involved a breakthrough method called “bootstrapping,” which allows indefinitely long sequences of encrypted computations to be performed.

FHE is part of a broader range of novel privacy technologies known as Privacy Enhancing Technologies, or PETs. These PETs provide a range of functionality, and it is typical for an application to leverage more than one PET to provide a complete solution for data privacy. FHE specifically provides a cryptographic assurance for the protection of data in use, reducing the complex and incomplete set of defences that are used to protect contemporary data processing systems to a single strong point of defence. Other PETs, such as differential privacy, can supplement FHE by providing (e.g) statistical guarantees of privacy in the outputs of a computation, such as statistical reporting.

FHE aligns with privacy regulations like General Data Protection Regulation (GDPR) and Health Insurance Portability and Accountability Act (HIPAA) by ensuring data privacy even during processing. Specific guidance on the usage of FHE and other Privacy Enhancing Technologies may vary by jurisdiction.

The adoption of FHE faces several challenges. First, relative to plaintext processing, the computing overhead of calculations performed in FHE is significantly higher.. Second, working directly with FHE requires specialised knowledge and infrastructure to implement efficient and secure applications. Lastly, FHE technology is a relatively new cryptographic technology with research and standardisation efforts underway to provide users with consistent best practices.

FHE can be performed under both symmetric-key and public-key implementations to support a range of applications. In all existing FHE schemes that have been subjected to extensive and open cryptanalysis, the cryptographic problem that secures the data is the “Learning with Errors” problem, a post-quantum resistant cryptographic primitive that also secures the next-generation cryptographic algorithms for secure key exchange over the internet that have been standardised by NIST.

In traditional encryption, data must be decrypted before any computation can occur, exposing it to potential risks. FHE allows computations directly on encrypted data, enabling secure data processing without revealing the data itself. FHE also belongs to a family of cryptographic technologies that are known to be resistant to quantum computing threats.

By design, FHE allows for arbitrary computing over encrypted data. However, some operations and algorithms are harder to implement under FHE than others. Contemporary tools for working with FHE range from complete software solutions that implement specific functionality through to powerful encrypted circuit compilers that enable new applications. These tools significantly expand on the capabilities and functionality of FHE by capturing expert knowledge in FHE optimisation and algorithm execution.