FHE: The Future Path of Privacy Computing

FHE( Fully Homomorphic Encryption ) is an advanced encryption technology that allows direct computation on encrypted data, thereby processing data while protecting privacy. FHE has multiple potential application scenarios, especially in the field of data processing and analysis that require privacy protection, such as finance, healthcare, cloud computing, machine learning, voting systems, the Internet of Things, and blockchain privacy protection. However, the commercialization of FHE still requires time, with the main obstacles being the enormous computational and memory overhead brought by its algorithms, as well as poor scalability.

Basic Principles

The core of FHE is to perform computations on encrypted data and obtain encrypted results that are consistent with the plaintext computation results after decryption. To achieve this goal, FHE uses polynomials to obscure the original information, as polynomials can be transformed into linear algebra problems and vector computation problems, facilitating optimizations such as parallel computing on modern computers.

The basic encryption process of FHE includes:

• Choose a key polynomial
• Generate random polynomial
• Generate small "error" polynomials
• Combine the plaintext with the above polynomial for encryption

To prevent decryption through repetitive analysis, FHE introduces noise. However, the noise accumulates during the computation process and may eventually lead to an inability to decrypt correctly. To address this issue, FHE employs the following techniques:

• Key switching: compresses the ciphertext size but introduces a small amount of noise
• Modulus Switching: Reduce noise by switching the modulus
• Bootstrap: Reset the noise to the initial level, but the computational overhead is significant.

Currently, mainstream FHE schemes are based on Bootstrap technology, including BGV, BFV, CKKS, and so on. These schemes have their advantages in arithmetic circuits and Boolean circuits.

Challenges Facing FHE

The biggest challenge of FHE lies in its enormous computational overhead. Compared to ordinary computation, the same computation in FHE versions may require up to 500 million times the computational resources. To improve the performance of FHE, the US DARPA has launched the Dprive program, aiming to increase the FHE computation speed to 1/10th of ordinary computation. The program mainly focuses on the following aspects:

• Increase processor word length
• Build dedicated ASIC processors
• Build MIMD parallel architecture

Although the Dprive project is about to expire, it seems to be progressing slowly and has not yet reached its expected goals. Similar to ZK technology, the implementation of FHE is also facing hardware bottlenecks.

Nevertheless, in the long run, FHE technology still has unique value, particularly in protecting sensitive data privacy. For critical sensitive data in fields such as military, medical, and financial, FHE can unleash the potential of technologies like AI while protecting privacy. This security is especially important in the post-quantum era.

The Combination of Blockchain

In the blockchain field, FHE is mainly used to protect data privacy, with application directions including on-chain privacy, AI training data privacy, on-chain voting privacy, privacy transaction reviews, and more. FHE is also considered one of the potential solutions to address the on-chain MEV problem.

However, FHE also faces some challenges. Fully encrypted transactions may eliminate the positive effects brought by MEV. In addition, running FHE on the virtual machine will significantly increase node requirements and greatly reduce network throughput.

Main Projects

The main projects in the field of FHE currently include:

• Zama: Built a relatively complete FHE development stack based on the TFHE scheme.
• Fhenix: Building a privacy-first Optimism Layer 2
• Privasea: Using FHE for LLM data operations
• Inco Network: Building FHE Layer 1
• Arcium: Integrating cryptographic technologies such as FHE, MPC, and ZK
• Mind Network: Solving consensus layer security issues through the FHE subnet architecture
• Octra: Implementing FHE using hypergraph technology to build a new smart contract language and consensus protocol.

Future Outlook

FHE technology is still in the early stages, and its development status is not as advanced as that of ZK technology. The main limiting factors include high costs, significant engineering challenges, and unclear commercial prospects. As the attention of crypto VCs towards FHE increases, it is expected that more funds and projects will enter this field.

The implementation of FHE chips is one of the key prerequisites for its commercialization. Currently, several manufacturers, including Intel, Chain Reaction, and Optalysys, are exploring this direction. Despite facing many technical hurdles, FHE, as a technology with great prospects and clear demand, is expected to bring profound changes in industries such as defense, finance, and healthcare, unlocking the immense potential of combining privacy data with future quantum algorithms.