Trustless: EigenLayer AVS will take the lead in detonating the Web3 privacy computing track with "cheap security".

Original author: @Web3 Mario

Introduction: EigenLayer AVS has been online for a while, in addition to its official guidance of EigenDA and Layer 2 and other related use cases, the author found a very interesting phenomenon, that is, EigenLayer AVS seems to be very attractive to projects in the privacy computing track, in the 9 AVS that have been launched, three belong to this track, including two ZK coprocessor projects Brevis and Lagrange, and Automata, a trusted execution environment project. Therefore, we decided to investigate in detail to find out what EigenLayer AVS means for related products and how to develop them in the future.

The attraction of “cheap security” is the key to the success or failure of the EigenLayer AVS ecosystem

With TVL officially exceeding $15 billion, EigenLayer has had a very fantastic start, of course, I think most of the funds are still aimed at obtaining potential Airdrop benefits, but this undoubtedly lays a solid foundation for EigenLayer to enter the next stage, and the key to the next stage is the success or failure of the AVS ecosystem, because the scale of AVS’s fee income determines the timing of EigenLayer’s transition from the subsidy period to the mature period.

There have been longer articles about the technical details of EigenLayer, so I won’t repeat them here. To put it simply, EigenLayer creates an inexpensive Consensus layer protocol by reusing the Consensus power of Ethereum Pos, also known as Restaking. First of all, I would like to explore the core values of EigenLayer, in my opinion, there are three main core values of EigenLayer:

*** Decouple the Consensus layer from the execution layer to better handle large-scale or expensive data processing and Consensus: Mainstream Blockchain protocol is generally considered to be a solution with high execution costs but low execution efficiency. We know that the Blockchain-based execution environment usually uses market machines to adjust the allocation of Node computing resources, that is, the higher bidders are given priority to be executed, and the executors are in a competitive relationship, and when the demand is high, the fair price will continue to rise, and the execution cost will inevitably go up.Block short The low execution efficiency stems from the fact that Blockchain technology is designed to be a settlement system for electronic money, and the processing of transaction data is time-sensitive. Therefore, the execution layer has to be designed in a serial way, which makes it inefficient when dealing with most of the time-insensitive scenarios, such as Social Web, AI training and other scenarios.

Drop the consensus layer from the execution layer, on the one hand, allows application developers to design a dedicated execution environment, commonly known as an AppChain or Layer 3, etc., so that its users can get rid of the competition with users of other applications and reduce the cost of use, on the other hand, it can allow developers to develop a more suitable execution layer according to different application scenarios and improve execution efficiency.

* Consensus-as-a-Service, fully exploiting the potential demand of the market by productizing or resourcing Consensus: I want to experience Layer 1 People who have been baptized in the era of hundreds of schools of thought will have a unified sigh, the big waves wash the sand, the establishment of the consensus layer is usually expensive and difficult, each family in order to maintain their own Consensus security guarantee, may be computing power or staked funds, before generating enough profitability are in the subsidy stage, and the cost is not high, usually the subject matter of the subsidy is the Token income obtained from mining. Only a small number of successful protocol can successfully rely on their own revenue capacity, i.e., fee income, to maintain sufficient Consensus capacity. For example, the shift in Ethereum’s economic model. This high start-up cost discourages long innovative applications because the cost of setting up an execution environment for their own application, or building their own AppChain, is too high and they face great risks. This makes the Matthew effect of the Web3 industry very obvious, and the evolution of the current Web3 technical solutions has been basically engulfed by the technical route of Ethereum.

By servitizing or productizing Consensus, innovative applications have another option, that is, to purchase Consensus services according to demand. As a simple example, for an innovative application, assuming that the entire application hosts $1 million in funds upfront, this means that as long as more than $1 million of PoS consensus is purchased, the security of its execution environment can be guaranteed, because the economic cost of evil is negative. With the development of applications, Consensus services can be purchased quantitatively flexibly. This drop the start-up costs of innovative applications, drop their risks, and exploit the full potential of the market.

* Cheap source of Consensus: The last point is that the Consensus source of EigenLayer uses the reuse of Ethereum’s PoS funds, which means that for PoS stake that can only capture one layer of income, participating in EigenLayer can long one layer of income, which allows EigenLayer to cleverly change the competitive relationship between itself and the industry leader Ethereum to a symbiotic relationship, drop the cost of attracting Consensus funds. This also makes it more advantageous than other protocols in terms of pricing, such as the consensus purchase fee of the AVS protocol, making it more attractive for innovative applications

The above three points make EigenLayer provide a source of “cheaper security” for Web3 applications compared to other Web3 execution environments, making it have lower execution costs, better scalability, and a more flexible business model. Therefore, I believe that the key to the vitality of the EigenLayer AVS ecosystem lies in whether Web3 applications can be impressed by this cheap security and migrate to the ecosystem in large quantities.

The cost of use is the fundamental reason that restricts the development of the Web3 privacy computing track

After discussing the core value of EigenLayer, let’s take a look at the dilemma of the Web3 privacy computing track. The author is not an expert in this field, so I focused on the current situation of the tracks related to privacy-preserving computing in AVS that are currently online. The so-called ZK coprocessor, I believe that most long Cryptography products that make use of zk-SNARKs Algorithm face the same dilemma, that is, the high cost of use hinders the promotion of use cases.

As the name suggests, the original intention of the related products of this track is to use zk-SNARKs Algorithm to provide coprocessor services for the current mainstream Blockchain system, so that it can offload complex and expensive computing operations to off-chain execution, and ensure the correctness of the execution results through zk-SNARKs. The most classic example of this modular idea is the relationship between the CPU and the GPU. Parallel computing operations such as image processing and AI training, which are not good at CPU architecture, are handled by another independent module, the GPU, to improve execution efficiency.

The technical architecture of a classic ZK coprocessor project is basically as follows, which is the simplified technical architecture of Axiom, which is one of the leaders of the track. To put it simply, when a user has a demand for a complex computation, you can use Axiom’s off-chain service to calculate the result and generate the relevant ZK Proof proof, and then Axiom will use the result and proof as parameters to call Axiom’s on-chain verification contract, which relies on the execution result, execution proof, and the key block information of the whole chain provided by Axiom officially to the on-chain, such as transaction merkle root (the process of maintaining the key information of the whole chain is also Trustless), and the correctness of the results is verified Algorithm verified by the on-chain. When the verification is passed, the result will be notified to the target contract through the pullback function to trigger subsequent operations.

It is generally believed that the process of proof generation is a computationally intensive operation, and the verification of proof is relatively light, through Axiom’s documentation, we can know that an on-chain ZK Proof verification operation needs to set roughly the verification gas fee is 420000, which means that assuming the gas price is 10 Gwei, the user needs to pay 0.0042 ETH verification cost for this. Assuming the market price of ETH is $3000, the cost is around $12. This cost is still too high for ordinary C-end users, which greatly limits the construction of potential use scenarios for this product.

Referring to a use case often advertised by a ZK coprocessor project, the Uniswap VIP program, that is, Uniswap can set up a loyalty program similar to CEX for its traders through the ZK coprocessor, and the trader’s cumulative volume in the past period of time reaches a certain magnitude, protocol the trader’s commission will be rebated or reduced. Considering that the calculation of cumulative volume is a complex operation, Uniswap can use the ZK coprocessor scheme to offload the computation to the off-chain, which drop the computational cost and avoids large-scale modification of the on-chain protocol.

You can simply calculate an account, assuming that Uniswap has set up a VIP activity that can enjoy full fee waiver as long as you can prove that your cumulative trading volume in the past 1 month exceeds $1,000,000, a trader chooses Uniswap 0.01% fee pool to trade, and when the user’s single trading volume is $100,000, the fee is $10, but the cost of verification is $ 12), which hits the user’s motivation to participate in this service, raises the threshold for activity participation, and in the end can only be more beneficial to Whale.

Similar cases should not be difficult to find in related ZK-only products, the use cases and technical architecture are great, but I think the cost of use is the core constraint that hinders the expansion of the use cases of related products.

From the transformation of Brevis, we can see the siphoning effect of EigenLayer’s “cheap security” on related products

So let’s take a look at how Brevis, one of the first to launch AVS, was affected by EigenLayer, and I hope to show that EigenLayer has a clear attraction to related cryptography products with its “cheap security”.

Brevis itself is positioned as a ZK coprocessor, and at the beginning of its launch in early 2023, it is still positioned as a full-chain data computing and verification platform, of course, this is no different from the essence of the ZK coprocessor, but the latter is cooler. For a long time, Brevis operated using the so-called “Pure-ZK” approach described above. This makes it difficult to promote use cases, and in a blog post on April 11, it announced a partnership with EigenLayer and a new “cryptoeconomics + ZK proof” solution, Brevis coChain. In this scenario, the validation layer sinks from the Ethereum Mainnet into a coChain maintained by AVS.

When the user generates a computational requirement, the result is calculated through the client circuit and the relevant ZK Proof is generated, and the computation request is sent to Brevis coChain through the on-chain smart contracts, and after listening to the request, AVS verifies the correctness of the calculation, and after passing it, the relevant data is packaged and sent to Ethereum Mainnet for some compression, and asserts the correctness of the result. In the coming period, as with other “optimistic verification” schemes, the challenger will enter a challenge period, where the challenger can dispute an outcome and fight for the slashing of the perpetrator by submitting the corresponding ZK fraud proof. After the slashing period has passed, AVS will use the callback of the target contract through the on-chain contract to complete the subsequent operation. Considering most of the issues of privacy-preserving computing, I would like to call this approach “optimistic trustless” when considering how to trustless through mathematics.

Similarly, Lagrange and Automata must have gone through the same journey before turning to a trustless solution that leverages AVS’s optimism. The benefit of this approach is that the cost of validation is greatly dropped. Because there is no longer a need for expensive on-chain verification computations in the process of achieving correct results, optimistic trust in the processing results of EigenLayer’s consensus layer and the security brought by ZK fraud proof. Of course, shifting from trust in mathematics to trust in humanity is bound to be challenged in the Web3 space. But I think it’s an acceptable result compared to the practicality it brings. Moreover, this solution will effectively break the constraints of verification costs on the promotion of use scenarios, and I believe that there will be long and more interesting products launched in the long .

This scheme also forms a demonstration effect for other privacy computing track products, considering that the track is still in the blue ocean stage, compared with the highly competitive rollup related track, it should be more conducive to the promotion of the new paradigm, I believe that the AVS ecosystem will be the first to usher in the outbreak of the privacy computing track, because the author is not related to the direction of cryptography, it is inevitable that there will be mistakes in the writing process, and I hope that experts will correct it.