Solution to MEV is simple:
* every user has their own block producer and makes their own blocks
Thus every user can guarantee that there is no undesirable MEV when they submit a tx. This is just impossible. Supporting tens of thousands of concurrent block producers is probably impossible, and sub 100ms slots with world wide replication is impossible.
Even if the network can clear 10 blocks per second or 800k blocks per day, the UX for a user that has to wait for half a day between blocks is not going to work for any interesting usecase. So this idea is a non starter.
So what is the next best thing?
Maximize the competition between existing block producers, and allow users to exclusively pick which block producer they want to allow to add their txs.
The way to maximize competition is still an engineering problem, it's just not impossible. The more block producers per second there are scheduled, the less time a user would need to wait for the best offer to include the user txs. Lower block times, concurrent leaders, faster rotation, etc... all improve up on this.
It's possible that worst case mev extracting block producer can offer a rebate to all users and still attract transaction flow.
This is up to the market to decide, but what is important is that the users willingly submitted their txs to this type of block auction.
So for this to work, some services needs to monitor and score validators on their behavior, and surface this information to the users. The RPC infra that deals with user txs has to correctly send the txs to only the users intended block producers. Potentially, the TXS themselves can be signed such that their are only valid in specific slots.
* every user has their own block producer and makes their own blocks
Thus every user can guarantee that there is no undesirable MEV when they submit a tx. This is just impossible. Supporting tens of thousands of concurrent block producers is probably impossible, and sub 100ms slots with world wide replication is impossible.
Even if the network can clear 10 blocks per second or 800k blocks per day, the UX for a user that has to wait for half a day between blocks is not going to work for any interesting usecase. So this idea is a non starter.
So what is the next best thing?
Maximize the competition between existing block producers, and allow users to exclusively pick which block producer they want to allow to add their txs.
The way to maximize competition is still an engineering problem, it's just not impossible. The more block producers per second there are scheduled, the less time a user would need to wait for the best offer to include the user txs. Lower block times, concurrent leaders, faster rotation, etc... all improve up on this.
It's possible that worst case mev extracting block producer can offer a rebate to all users and still attract transaction flow.
This is up to the market to decide, but what is important is that the users willingly submitted their txs to this type of block auction.
So for this to work, some services needs to monitor and score validators on their behavior, and surface this information to the users. The RPC infra that deals with user txs has to correctly send the txs to only the users intended block producers. Potentially, the TXS themselves can be signed such that their are only valid in specific slots.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
