Source: Cointelegraph Magazine, translated by imToken.

Editor's Note: Ethereum is moving towards a new era of scaling with 10,000 TPS, and zero-knowledge proof (ZK) technology is becoming a key driver. This article will focus on the technical difficulties of real-time proving, the participation logic of Provers, security challenges in the L1 switching process, and how 'native Rollup' will become the ultimate form of ZK scaling.

If ZKification is the starting point for the technological reconstruction of Ethereum, then 'real-time proving' and 'native Rollup' are the core landing segments of this scaling revolution.

In this article, we will continue to explore how to achieve 12-second level ZK real-time proving on the Ethereum mainnet, what the hardware thresholds and incentive mechanisms are for becoming a Prover, and how native Rollup will rewrite the landscape of Ethereum L2.

Real-time proving: The key piece in Ethereum scaling.

On the roadmap for Ethereum's goal of achieving 10,000 TPS, there is an indispensable technical breakthrough: real-time proving.

Succinct co-founder Uma Roy explained it this way: 'Real-time proving refers to the ability to complete the generation process of ZK proof for a block on the Ethereum mainnet in less than 12 seconds.'

What does this mean? Once real-time proving is achieved, Ethereum will be able to incorporate its block validation logic into the protocol itself, and nearly 'arbitrarily' increase the Gas limit without sacrificing verifiability, thereby enabling large-scale scaling of L1 (Note: the generation time for each block on the Ethereum mainnet is 12 seconds, so 'real-time' means completing the proof within each block cycle).

However, to achieve real-time proving, relying solely on zkVM technology is not enough; changes to the Ethereum protocol layer are also required.

Ladislaus from the Ethereum Foundation points out that a key mechanism is expected to be introduced in the Glamsterdam upgrade next year—'decoupling block validation from immediate execution', which will provide Provers with more ample time to generate zkEVM proofs within a complete slot, thus achieving true real-time processing.

In terms of technical implementation, Succinct has released its latest SP1 Hypercube zkVM, capable of generating proofs in real-time for 93% of the 10,000 mainnet blocks under a cluster of 200 GPUs.

Roy stated that they are confident in raising the success rate to 99% by the end of this year. Although some difficult-to-handle blocks may still lead to a very small number of blocks being unable to generate proofs in time, the protocol design has already considered fault tolerance mechanisms, such as allowing the skipping of that block and proceeding to the next block for processing.

Furthermore, Ethereum is also considering reducing block times from 12 seconds to 6 seconds (as another potential proposal for Glamsterdam), which would significantly enhance user experience and transaction confirmation speed, but this also places additional pressure on ZK Provers—doubling the difficulty of the task for Provers.

However, Roy is not worried, as ZK technology can improve performance tenfold each year; even if block times are halved, it can still cope.

In June, Linea also announced that 100% of on-chain activities on its network could be covered by ZK proofs. Although Linea's current TPS is only 2, this is not a performance limitation but is constrained by usage demand.

It is worth noting that Linea has an inter-block interval of just 2 seconds, with ZK proofs uploaded to Ethereum L1 for verification via smart contracts. This model may well be a precursor to the future 'ZKification' of the mainnet.

Is the hardware threshold for Ethereum ZK provers high?

To generate ZK proofs in real time, powerful computing resources are essential.

The preliminary technical goal set by the Ethereum Foundation for Provers is to keep hardware costs below $100,000, with power consumption below 10 kilowatts, roughly equivalent to the power consumption level of a Tesla Powerwall home battery.

This figure does not sound 'lightweight'. Ethereum critic Justin Bons (founder of Cyber Capital) called it 'mad hardware requirements far exceeding Solana validator nodes', but this actually confuses two completely different roles.

Ladislaus from the Ethereum Foundation protocol coordination team points out that the responsibilities of Provers and Validators are different and cannot be confused—Validators run nodes and participate in consensus, while the task of Provers is to generate ZK proofs. Once the ZK proof for a transaction is correctly generated, the network only needs to verify whether the proof is correct and does not need to re-execute the transaction.

It is precisely for this reason that Ladislaus expresses optimism, stating, "As long as an honest prover that meets the hardware conditions can be found, Ethereum can continue to operate securely. We intentionally set the threshold below that of data centers, and even individuals with technical capabilities can run Prover at home, even if they are not large institutions or data centers."

Currently, this hardware configuration of $100,000 is just an initial target, and Ethereum Foundation researcher Sophia Gold expects that mainstream Provers will meet this target by the Devconnect developer conference in Argentina this November.

Succinct co-founder Roy expects that by early next year, GPU requirements can be reduced to around 16 graphics cards, and the total cost will be controlled between $10,000 and $30,000.

Meanwhile, Succinct has built a decentralized network consisting of "hundreds of Provers" on the test network, generating millions of proofs cumulatively.

The core logic of this system is competitive proving, where all Provers participate in bidding, selecting one winner each round to execute the zk proof, aiming for the participant with shorter time and lower cost to win, thus forming a computational power bidding mechanism.

This means that in the ZK-driven future of Ethereum, the spirit of mining will re-emerge in another form—only the role has changed from computing blocks to computing proofs.

Mainnet switch to ZK architecture: A high-difficulty system migration.

Switching the Ethereum L1 mainnet to a zero-knowledge proof (ZK) architecture is yet another technical challenge of almost the same level, following the transition from proof of work (PoW) to proof of stake (PoS) in 2022. The entire process not only requires a reconstruction of the protocol layer but also must carefully consider various potential edge scenarios and security risks to prevent network disruptions.

At the EthProofs conference in July this year, researcher Justin Drake mentioned several potential risk hazards. For example, malicious attackers may insert so-called 'prover killers' into blocks, leading to the failure of the entire network verification mechanism; or a sudden drop in network activity could result in transaction fee revenue being insufficient to cover the cost of generating ZK proofs, thus affecting network sustainability.

Ladislaus from the Ethereum Foundation protocol coordination team stated that the entire transition process may take several years, especially with a focus on security risks. The ZK virtual machine (zkVM), as a complex technology still in its early stages, is likely to have various vulnerabilities. However, as the ecosystem matures, we can gradually improve its feasibility and robustness on Ethereum L1 through the introduction of proof diversity, the improvement of incentive mechanisms, and formal verification.

Meanwhile, Ethereum also plans a fundamental architectural overhaul of its consensus layer, aiming to build a new structure called 'Beam Chain' that is friendly to ZK optimization from the design phase. Drake even stated that in the future, all of Ethereum’s data verification work could be completed on the CPU of an ordinary laptop.

Mainnet 'Snarkification': Native Rollup is coming.

At the same time as integrating zkEVM into the Ethereum mainnet, another long-term vision is gradually emerging: native Rollup.

Currently, Rollups (whether Optimistic or ZK type) adopt independent proof systems, with their security relying on their own validators or ordering mechanisms, which involves a certain level of trust assumption between them and the Ethereum mainnet.

The vision of 'native Rollup' is entirely different—by integrating zkEVM into the mainnet, allowing Ethereum L1 validators to directly verify the Rollup's state transition proofs, thus achieving a truly mainnet-validated, mainnet-secured L2.

This requires adding a key piece of code 'execute precompile' to the Ethereum L1 client, allowing validators to directly verify the ZK state transition proofs generated by L2. As Ethereum Foundation protocol coordinator Ladislaus said, 'L1 validators will consume the execution proofs from these Rollups and verify their correctness.'

In other words, if native Rollup becomes a reality, then in the future, whether a transaction occurs on L1 or on a native Rollup, the final settlement and security will be guaranteed by the same group of Ethereum validators, and the level of trust will be entirely equivalent.

This means that depositing $10 million into a native Rollup will have the same level of security as directly depositing into the Ethereum mainnet.

Declan Fox, the project lead for Linea, stated that their long-term goal is to become a native Rollup. He believes this is an 'upgraded version' of the ETH 2.0 sharding solution—no longer rigidly operating 64 structurally identical shard chains, but building heterogeneous Rollup systems in a highly programmable and customizable manner to serve different scenarios and user needs.

Unlike the homogeneous sharding architecture of ETH 2.0 in the past, native Rollup can be heterogeneous, providing end users with a more diverse and differentiated application experience.

Although native Rollup has not yet been formally written into the Ethereum roadmap, with zkEVM officially starting and L1 architecture gradually being restructured to preset interfaces and precompiled logic, it has obviously become a foreseeable technical trend.

Ladislaus summarized, 'In integrating EVM Snarkification (i.e., integrating ZK proof capabilities) and advancing native Rollup, Ethereum has a high degree of technical synergy, as both share the underlying ZK technology stack.' Of course, this process still needs to be governed by the Ethereum community to form EIPs (Ethereum Improvement Proposals) and ultimately be deployed in a hard fork.

Optimistically, if everything goes smoothly, perhaps by the end of the year, relevant EIPs could be submitted and launched in the fork after the Glamsterdam upgrade.

However, this timeline still has a high degree of uncertainty and should be viewed with caution.