I’m Madhumitha Harishankar, CEO of Nume Crypto. Today, I’m excited to announce the testnet launch of the payments protocol that we’ve been building in stealth for over a year — the Nume Protocol. Nume disrupts existing scaling solutions in the Ethereum ecosystem by offering an entirely gas-abstracted infrastructure with a flat fee(taken in DAI) monetization model for all token and NFT activities, doing away with variable gasprice and percent fees. Let’s dive in.
We believe web3 will be brought to the masses on the most trusted, decentralized, and secure Layer-1 blockchain — Ethereum — notwithstanding its current scaling limitations. At 30TPS, variable and expensive gas costs are a deterrent even to early adopters and will be a massive barrier in onboarding the next billion into web3 — users who will not put up with confirmation delays, high gas fees, or even just variable fees.
The predominant approaches to trustlessly scaling Ethereum involve trading off between cost and settlement finality time, and the Nume protocol exists to bridge this gap. Our infrastructure scales supported operations without using zero-knowledge proofs while also providing instant settlement finality on-chain. Additionally, Nume can batch an unlimited number of transactions for a fixed on-chain commit cost, enabling dApps to hyper-scale and bring web2-like traffic to web3 at the lowest cost. Note: that Nume will not support smart-contract deployment at launch— it will support asset transfers and specific predefined operations (defined by Nume). In this aspect, it’s similar to zkSync Lite or Loopring .
Existing Approaches
We provide a brief (simplistic) overview of Optimistic Rollups (OR) and Zero-knowledge Rollups (ZKR) before diving further into our protocol. With OR, a sequencer batches transactions and sends the compressed transaction data and new state commitment to the OR contract on Ethereum. A challenge-response (C-R) period now commences, allowing users/validators of the network to discover and prove incorrect state transitions in this new state before it is finalized on Ethereum. While ORs offer fast transaction confirmation times (off-chain) and incur low costs, this C-R period, which usually lasts 7 days, delays settlement finality significantly (and undesirably).
In ZKRs, similar to ORs, the sequencer batches multiple transactions and computes the new state transition. However, this time the sequencer (or a verifier) also generates a zero-knowledge proof that cryptographically proves the state transitions to the new commitment. This proof is sent to the rollup contract on Ethereum, which instantly accepts the new commitment if the proof verifies. The drawback, however, is the substantially high resource cost incurred to compute zk proofs, which typically results in a higher gas cost passed onto users (compared to ORs).
Nume’s Approach
At Nume, we mitigate this cost-finality tradeoff by combining hardware enclaves with decentralized governance. We execute the rollup algorithm (that verifies transactions and computes the new state) in a hardware enclave, also called Trusted Execution Environment (TEE). The new state generated at the end of the computation is attested by the TEE and one of the attested measurements is a reference to the source code itself. The on-chain contract is configured to accept new state commitments as long as it is accompanied by an attestation from the correct source code. But who decides the correct source code? This is where decentralized governance comes in — while multisigs and DAOs typically govern on-chain code, in Nume’s case they also govern the off-chain code that runs inside the TEE. Any modifications to the off-chain rollup algorithm must be voted on via on-chain governance for the contract to start accepting commitments with attestations that point to the modified source.
While we discuss further details in a separate post, it is important to highlight the efficiency of Nume’s model. Given an authorized source code, state commitments are instantly accepted by the on-chain contract as long as the conditions on the attested values pass. Then users only need to watch the contract for any proposed updates to the rollup software. They can verify an update (whether to the onchain or offchain protocol code) themselves during the cooling period that the contract enforces, and, if unconvinced by the governance outcome on the update, exit their funds. Either way, users do not need to watch the chain constantly to verify the validity of every single posted settlement, which is a big win in usability.
Since the Nume protocol proves the off-chain computation that was performed to the contract without any computationally expensive zk proofs, it also incurs much lower gas costs than ZKRs. It brings the power of centralized computing solutions to blockchain scaling in a trustless manner without the cost limitations of ZK proving or the usability limitation of long finality times. It also enables an entirely new security paradigm for accessing off-chain services and data on-chain; by embedding such access inside the rollup software that executes in the off-chain TEE, we can offer foundationally more secure oracles (and other services). A deep dive on Nume’s security model, limitations of its current implementation, and the roadmap to full trustlessness and decentralization is covered here.
We can’t wait to bring Nume to you in its closed mainnet beta in early Q3 2023. If you like to be in great company and want access to our beta, let’s chat. Our docs for testnet are available here. We would love to hear your feedback — if we can make development easier or optimize the stack for your particular use case, please reach out via Twitter We’ll expand on our roadmap in a subsequent post. We are excited to dialogue with the community to inform our feature development roadmap so please reach out to chat! :)
