What are some solutions on Ethereum to mitigate MEV at the non-protocol layer?

Ethereum researcher Jason Chaskin discusses the MEV issue, focusing on two mainstream off-chain solutions at the non-protocol level: off-chain optimization and private memory pool Mempool. He introduces the implementation principles and explains why these mechanisms will not completely replace the public memory pool. Discussion

Solutions to Minimize MEV

Miner-extractable value (MEV) remains a significant issue on the Ethereum network, where front-running transactions by focusing on the mempool to extract value from miners is a prevalent phenomenon. Essentially, transactions conducted through the public mempool are susceptible to MEV, and Ethereum's 12-second block time creates a window that arbitrageurs find easily exploitable. Consequently, the Ethereum Foundation has identified the MEV problem as a major challenge to address on its development roadmap.

Aside from the Ethereum Foundation's PBS route, there are currently two primary market solutions beyond the Ethereum protocol layer:

• Off-chain optimization: Users sign intentions off-chain to have their transactions analyzed and broken down into multiple optimized transactions to reduce the profit margin for MEV.
• Private mempool: User transactions are directly sent to builders, bypassing the public mempool entirely, limiting the ability for searchers to profit from transactions.

The goal of both approaches is essentially the same: to avoid using public mempools to reduce the impact of monitored transactions being exploited for MEV.

Solana's MEV Moment: Time to Seriously Address MEV Issues

Off-chain Optimization

These solutions are typically applied to applications that are particularly vulnerable to MEV attacks, such as AMMs. Currently, 98.85% of MEV extracted on the Ethereum network comes from three AMMs: Uniswap V2 at 63.4%, Balancer V1 at 19.39%, and Uniswap V3 at 16.06%. The designs of these AMMs allow searchers to profit from slippage. When users submit transactions to the public mempool, searchers can manipulate orders to ensure users receive the worst price within specified limits and profit from the price difference.

The "MEV optimization applications" mentioned here specifically refer to DEX aggregators, such as 1inch, Matcha, or UniswapX. The principle behind them is essentially the same: users express transaction intentions off-chain and analyze and execute them on-chain in a way that minimizes slippage and MEV risks.

By using off-chain intentions to optimize transactions before they enter the public mempool, the MEV problem is alleviated.

There are two common practical methods for off-chain MEV optimization:

Smart Routers

The first method involves smart routers, where users submit intentions to servers—for example, exchanging 1 ETH for 3,500 USDC. The smart router analyzes the liquidity of multiple AMMs and designs the optimal trading route, possibly dispersing orders across various pools to obtain the best price. This minimizes potential slippage, making the transaction less profitable for searchers in terms of MEV.

Off-chain Order Matching

The second method is off-chain order matching, used by systems like CoW Swap. Decentralized solvers match buyers and sellers off-chain, then execute these matched orders on-chain in batches to save on gas costs.

Although some level of trust in off-chain components is required, DEX aggregators are increasingly popular, currently accounting for 22% of decentralized exchange trading volume.

Recommended Read: UniswapX Launched! Trading Aggregator 1inch and MEV-Protecting DEX CowSwap Crisis?
Reason for Recommendation: Essential for readers unfamiliar with UniswapX to quickly understand its operational mechanism and the introduction of Fillers and intent-oriented transactions.

Private Mempool

The second way to avoid public mempools is by using private transactions. Private transactions are sent directly to builders or builder groups without broadcasting them to the public mempool, with approximately 10% of transactions currently being privately sent.

Currently, there are two main methods of private transactions: private RPC and Order Flow Auctions (OFAs).

Private RPC

A typical example of private RPC is Flashbots Protect. Unlike broadcasting transactions to the public mempool, Flashbots Protect submits transactions directly to selected builders to prevent searchers from seeing these transactions. While builders may share limited transaction information with some searchers to increase efficiency, this information is insufficient for searchers to engage in front-running. Since its inception, Flashbots Protect has safeguarded over 13 million transactions and processed nearly $30 billion in transaction volume.

Order Flow Auctions (OFAs)

The second method of private transactions involves Order Flow Auctions (OFAs), which essentially add an auction feature on top of the concept of private RPC. The primary concept is to allow users to maximize their share of MEV through auction competition.

Users submit their orders or intentions, and MEV searchers compete to execute these orders. While searchers still aim to maximize their MEV strategy, the competition in the auction compels them to aggressively lower prices, ensuring that most MEV profits flow back to users. For instance, a user wishes to exchange 100 ETH for 350,000 USDC. If the transaction is directly submitted to the public mempool, MEV bots may manipulate the order leading to significant slippage, and the user may only receive 320,000 USDC. However, in OFAs, searchers compete to execute this transaction, potentially offering a higher amount back to the user, such as $29,000, allowing the user to receive 349,000 USDC, significantly mitigating the negative impact of MEV.

OFA is a relatively new concept still under development. Initially launched by the now-closed Rook in 2021, projects like UMA's Oval and Flashbots' MEV-Share currently provide services in this area.

UMA Introduces Anti-MEV Tool Oval to Reduce Protocol Losses Caused by Oracle Feeding Gaps

Private Mempool Will Not Replace Public Mempool

Bringing up an interesting question, if private mempools can maximize user rights, is there still a need for the public mempool? Chaskin believes that the public mempool still serves a crucial function that needs to be retained—censorship resistance.

The role of the public mempool will evolve, serving a specific but critical purpose: maximizing resistance to censorship.

Transactions sent via private RPC completely bypass the public mempool, meaning validators running MEV-Boost and outsourced block-building will never see these transactions. If these validators begin conducting censorship, similar to what most relayers are doing now, some transactions may fail to be successfully included on-chain.

Therefore, the public mempool will not disappear entirely. It will evolve into a niche tool for specific scenarios critical, such as transactions requiring resistance to censorship. A small portion of users will continue to utilize the public mempool, prioritizing on-chain transactions facing potential censorship risks over MEV minimization.