Documentation on the breakthrough hash function for ZKproofs & Verifiable Computation, co-authored by Dusk Network's Dmitry Khovratovich, is now live and public.
Earlier this year, we introduced the blockchain and zero-knowledge community to our new hash function, co-authored by Dusk Network’s lead cryptographer Dmitry Khovratovich: Reinforced Concrete. This development represents a massive leap in the space, greatly affecting the scalability and speed of zero-knowledge processes and verifiable computation. Reinforced Concrete opens the door to the creation of entirely new use-cases and makes proving more computationally efficient.
👉 Quick link: go straight to the Reinforced Concrete Paper
In this article, we’ll be providing brief context to the technical research paper, discussing the unique benefits of Reinforced Concrete, its performance compared to other hash functions, and its effect on Zero-Knowledge Cryptography.
Zero-Knowledge: A Brief Introduction
Zero-knowledge (ZK) technology forms the bedrock of Dusk Network and, we believe, the future of the financial landscape. The ability to create verifiable ‘proof’ of a transaction (or its individual parameters) without revealing the information of said transaction or data, is what makes ZK technology viable for many use-cases otherwise impossible to perform with traditional blockchain technology. Zero-knowledge (ZK) proof systems combine various cryptographic tools to create brand new applications. Across the globe many researchers and developers, including Dusk Network, are exploring the potential of zero-knowledge tech.
If you’d like to learn more about ZK technology and ZK Proofs, we highly recommend reading our Hackernoon Compendium articles.
A Leap Forward: The Improvements of Reinforced Concrete
Reinforced Concrete is a new hash function that greatly improves the speed and scalability of ZK Proofs: it is a fast, secure, and zero-knowledge proof-friendly hash function that is up to 17 times faster than Poseidon, the current ZK hash function standard. To fully grasp this leap in performance and analyze the benchmarks between these hash functions, we must first contextualize these numbers and understand the meaning of its variables.
It is well known that privacy in the blockchain space does not come for free: a money transaction in the totally private ZCash required almost a minute to assemble due to the usage of the fairly standard, but the zero-knowledge unfriendly hash function SHA-256. The only tradeoff so far was to use a specialized algebraic hash function which reduces the transaction time but greatly increases the time to just collect all transactions. Reinforced Concrete is the first and only function that is both fast in zero-knowledge and regular usage, thanks to its novel design.
Its other unique feature is the reliance on the much older security analysis, which places far more confidence in it. Speaking of mid-term future, Reinforced Concrete makes the so-called universal verifiable computation, when every computing is provable, far more practical.
As you can see in the benchmarks below, when compared to other major hash functions, previously used in the same scenarios, Reinforced Concrete stands head and shoulders above the competition, with the lowest times to complete a full hash. In regular computation is 5 to 17x faster than Poseidon, 100 to 300x than Rescue, 100x than the newest ZCash design Sinsemilla, and is within reach (30%) of the SHA-2 performance. In Zero-Knowledge scenario, its advantage is less noticeable but still unbeatable.
💡 To avoid getting too technical, BN (Barreto-Naehrig Curves), BLS (Boneh–Lynn–Shacham Curves), and ST (STimpact) are the mathematical curves by which performance is measured.
As this hash function is zero-knowledge friendly, these numbers are equally impressive when constructed in-circuit. With only 267 lookup constraints, and a lookup table of 1500 elements, Reinforced Concrete provides algebraic security similar to previously deployed well-known standard cryptographic functions, such as AES. As a result, we can reduce the constraint count of in-circuit hashing to reduce the computing times of contracts across the dusk stack.
Presenting The Research Paper In Full
With Dmitry Khovratovich’s knowledge and ability backed by Dusk Network, we’ve taken Zero Knowledge Proof technology to the next frontier; the utility of which we’re only now beginning to discover. Benchmarks have shown a technological leap set to change the possibilities and capabilities of Zero-Knowledge implementations. With the markedly improved performance Reinforced Concrete provides, it will serve as one of the pillars of Dusk Network technology and, we’re sure, much of the ZK industry. We highly recommend taking a look at the full Research Paper and our previous article on Reinforced Concrete to get a comprehensive overview of this breakthrough.
Without further ado, dive straight into the Reinforced Concrete Paper: