From MEV to Execution-as-a-Service
MEV recap
“A blockchain can be viewed as a timestamping service that allows for the ordering of data in a canonical way” — Paradigm
Data on a blockchain can only be changed by signed transactions. Therefore, the right to order transactions lies at the heart of blockchain economics. Indeed, the buzz word “MEV” (Miner-Extractable-Value or Maximal-Extractable-Value) is defined as the value that can be extracted from transaction inclusion, exclusion and ordering, by miners, validators, sequencers, builders, etc. For a more in-depth discussion of MEV per se, here’s an old article by me.
Think MEV from orderflow angle
Ethereum MEV is a huge cake — hundreds of thousands to millions of dollars profit every day. In the beginning it was monopolized by miners when transaction ordering was fully determined by PGA (Priority Gas Auction); then Flashbots came into play, formalizing MEV extraction by searchers and miners with sealed-bid auction.
A weird fact is, it’s users who actually initiate and own their transactions, but they don’t get to share any of the MEV generated from those transactions. Users have long been excluded from the MEV discussion.
The reason behind this is that users give their valuable orderflow to searchers for free. We simply send transactions to Ethereum public mempool via MetaMask default RPC (which is Infura) for searchers to play their Dark Forest game. In fact, Retail orderflow is valuable to market makers because of non-toxicity (in other words, being more naive than institutional takers. You can refer to this article for more info), who are willing to pay for it. In tradfi it’s called PFOF (Payment For Order Flow) and that’s how Robinhood made most of its money: by selling Robinhood users’ orders to market makers. Users get rebates too.
So, how can i sell my orderflow in Crypto?
PFOF has another name: Exclusive orderflow (EOF). And it exists in crypto too. For example, all Sushiswap orders are exclusively routed through (aka sold to) Manifold Finance, a private relay that’s unaccessible to public mempool searchers, who then extracts as much MEV as possible via backrunning, and split profits between themselves and Sushi, which is then shared among Sushi stakers. This relationship is exactly like Robinhood and its MMs — brokers (Robinhood/Sushi) give exclusive non-toxic orderflow to professional MMs (Manifold and its searchers) and share part of the profits generated. I call it the 2B solution because Manifold works with dapps.
Another example is Rook. Rook is what I call the 2C solution (although it provides services to dapps too): it works directly with users. If you want to swap 1M DAI for ETH on chain, you can submit your transaction through Rook, which is also a private relay protecting you from outside searchers. Searchers on Rook, who stake $ROOK for access to transactions, bid for the right to execute your transaction. Part of the bid then goes back to you.
Other than those two on Ethereum, probably the most classic instance of DeFi PFOF is DFlow, which started on Solana and recently launched an appchain.
MEV and EOF on the PoS Ethereum
EOF protocols began to receive magnitudes more attention after the Merge, and subsequent Proposer-Builder Separation implemented by Flashbots’ MEV-Boost client. MEV-Boost implements sealed-bid auction to the validator layer: a new role called builders collect transactions, build the whole block, and only provide the block header and a tip to ETH validators (also called block proposer). Proposers don’t see the content of the blocks, therefore cannot extract MEV by themselves. Instead they simply propose the block with the highest tip, adding it to Ethereum blockchain. MEV-Boost and PBS, which will be enshrined to Ethereum in the future, essentially isolate validators from MEV extraction, preventing potential harm from MEV to Ethereum consensus. Right now, nearly 90% of all Ethereum blocks are produced by MEV-Boost clients.
What does this have to do with EOF? MEV is now highly professionalized and concentrated to block builders. In order to build the most lucrative block, builders need access to the most lucrative transactions — where EOF comes into play. Those who control orderflow control Ethereum block production.
Centralization is the end of the world…or is it?
And this is the reason why many are worrying about a potential MEV centralization: for example MetaMask could set up their own transaction relay, searcher and builder, and monopolize block building with MetaMask orderflow, which is 70% of all Ethereum transactions. Even searchers in public mempool have to submit their bundles to the MetaMask builder whose blocks get proposed all the time (because they are simply more lucrative than other builders’ blocks thus can pay higher tips to validators). By that time, no party would have enough bargaining power against MetaMask, whose monopoly makes everyone else worse off.
It’s such a gloomy outlook. But if we push the thought experiment a couple steps further, we will see that this type of vertical-integration-based monopoly is easier to break than it seems, because crypto users have a stronger ownership over their transactions than in tradifi. Let’s say if MetaMask obtains the monopolistic position, and starts to squeeze everyone else — validators get less tips, users get less rebates — what will happen next? Users will surely be unhappy. They will look for other channels to sell their orders for a higher price. DEXs have a natural advantage to satisfy this need, since the majority of MEV arises from DEX trading activities. If Uniswap, Sushiswap and Balancer forms an association and build their own transaction relay (and searcher+builder team) and offer users with better pricing, a huge chunk of valuable orderflow will leave MetaMask and go to that “DEX relay” — switching cost is relatively low, since all you need to do is signing a message. The MetaMask builder now faces a strong contender: the DEX builder, who is supercharged by DEX orderflow to build lucrative blocks and get them proposed. Competition between both brings more value sharing to validators and users.
Ownership gives birth to EaaS
Switching your stock broker is a costly activity, but switching RPC is only one-click. It again shows that blockchain gives users true ownership. With that in mind, applications — MetaMask, Uniswap, etc. — are most likely to go for adequate competition from day one. Execution-as-a-service (EaaS) thus becomes plausible. EaaS looks just like a generalized Rook: instead of just swap transactions, all transactions can be sent to a transaction relay, where searchers bid for the right to execute, part of which goes back to users, aka transaction originators.
See this is a searcher-layer full competition, as users don’t really care about which builder searchers are sending bundles to with how much tip. Even if we have a dominant builder who builds and wins 90% of all blocks, it still doesn’t necessarily mean lower return to users, as long as it still relies on searchers to provide bundled transactions. The key is to avoid builder’s control over orderflow, which according to our previous deduction, is economically implausible.
Conflicts only happen when the dominant builder starts its own searcher and discriminates against outside searchers — which will lead to searchers going to other builders, reducing orderflow going to the builder and harming the builder itself eventually.
Such an EaaS neutral relay has the potential to fundamentally change how people transact on-chain. By EaaS, users can now cut a bigger share in the market surplus that’s previously monopolized by institutions. Rook is building a neutral relay and it seems to be a transaction relay. Something worth following.
What comes with EaaS is MEV redefined. Today, MEV only seems to derive from most straightforward DeFi activities: sandwiching/backrunning transactions, liquidation, arbitrage, etc. But we are seeing MEV expanding its boundary. In theory, every transaction carries a certain amount of MEV. More people, facilitated by more available information, begin to extract more uncertain (but potentially much bigger) value. EaaS makes sure that more complex MEV strategies can be built, and users will benefit from the ever-growing MEV.
