Fhenix, a potential layer 2 network that aims to bring fully homomorphic encryption (FHE) to smart contracts, has unveiled a technical white paper for its first FHE rollup.
FHE itself is a form of encryption that guarantees the confidentiality of the data even when used, for example in smart contracts, because data can remain encrypted and processed without ever having to be decrypted.
This allows Ethereum and other EVM networks to process information privately, with potential applications related to private voting and confidential gaming, such as poker.
Read more: Encryption startup Fhenix raises $7 million for private computing
Unlike zero-knowledge (ZK) technology, FHE is completely private, Fhenix CEO Guy Itzhaki told Blockworks.
“With ZK, there is an entity called a prover that actually sees the information,” Itzhaki explains, “so the confidentiality comes from the fact that this prover is placed on the user side and produces the ZK proofs.”
One limitation is that it excludes use cases that require calculations based on data coming from multiple users, he said.
This is not the case for FHE-based systems, Itzahki said. With FHE, data and computation are performed on-chain, without the need for an processor or other entity.
“Our overall vision at Fhenix is to create an FHE stack that allows developers to build applications with data confidentiality, and allows users to benefit from these applications by increasing their security and privacy levels,” he said.
FHE itself goes beyond just protecting privacy, Guy Zyskind, the founder of SCRT Labs (the team behind the Secret Network) told Blockworks.
“This isn’t just about protecting people’s privacy; the lack of on-chain encryption is why we are very limited today in the applications we can build on blockchains. For example, consider creating an on-chain poker game. Without encryption, players’ hands would be visible to everyone, making the game unplayable,” he said.
Zyskind notes that existing privacy-focused blockchains rely heavily on zero-knowledge proofs, which fall short when it comes to integrating private data from multiple sources, such as in an on-chain poker game or a sealed-bid auction contract.
“Zero-knowledge proofs are more suitable for single-user scenarios, such as verifying age without revealing the exact date of birth. In these cases, zk-proofs are effective, but insufficient for complex multi-party private data integration, where FHE rollups become essential,” he said.
Using an optimistic rollup structure
The FHE rollup white paper outlines the use of an optimistic rollup structure, a decision, Itzahki notes, that allows the rollup to be both scalable and private.
While the use of a zk-rollup architecture isn’t completely out of the question, Itzahki explains that it would be difficult to create a rollup solution that combines two very powerful cryptographic systems.
“At the moment, the implementation of FHE via is optimistic [rollup structures] is much more efficient,” he says.
One of the key breakthroughs for Fhenix was coming up with an effective fraud resistance mechanism when using an optimistic rollup.
Read more: So your layer 2 is ‘secured by Ethereum’ – what does that mean?
Itzahki explains that the team has managed to introduce a tamper-proof mechanism using web assembly, a low-level computer language, that allows fraud verification to be performed in the same way as if the data were not encrypted.
“Rather than [running] On the native EVM itself, that mechanism allows us to check if the web assembly code is a change, and lets us know if there is a fraudulent transaction,” he said.
Modular design
The FHE package will be modular by design, meaning it will draw from several components to create the blockchain.
“We have a separation between the sequencer, validator and the data availability layer,” Itzahki said. “The data availability layer is usually Ethereum, then we at Fhenix will handle the sequencer and manage the execution, the validator and the prover.”
A modular design allows cryptographic innovations like FHE to be implemented more efficiently, allowing it to focus on its strengths where necessary.
“We don’t necessarily have to have that [FHE] through all components. That is why we opt for a modular design,” Itzahki adds.
Kyle Samani, managing partner at Multicoin Capital, an investor in Fhenix, said the rollup of the protocol is a first layer of its kind that will bring FHE onto the chain in a useful way.
“The white paper is a critical first step in helping developers realize that L2s can be used for more than just scaling. They can now be used to extend the functionality of applications into new design spaces,” said Samani.
Fhenix plans to launch the testnet of its first FHE rollup in the first quarter of 2024. It is unclear whether the rollup stack will be built open-source, although testing will be available for all interested projects.