How to use ZK and VDF to realize the idea of "privacy mempool"?

Most of the current L2 Sequencers basically adopt the “first-in, first-out” (FIFS) transaction ordering method to protect users from MEV, but this also weakens the value of blocks.

And through this split Blockspace solution, we can have both fish and bear’s paw.

The specific process is: the user uses a “time puzzle” to encrypt his transaction, and at the same time calculates a “zk proof” to prove that the time puzzle “has a solution”, and then uses the “time puzzle” and the corresponding “Zk proof” Proof" and sent to “Sequencer”.

After the Sequencer receives the “encrypted transaction”:

1. Verify whether the “zk proof” is valid. If it proves valid, it means that this “time puzzle” can be solved after a period of calculation;
2. Put it in “Top Blockspace” and give the “Order Committee” in the block where the transaction is located;
3. Sequencer will calculate the “time puzzle” for a period of time, and finally come up with an answer;
4. After getting the answer, the Sequencer can decrypt the user’s “encrypted transaction” and get the “original transaction” data;
5. After the Sequencer fills up the “Top Blockspace”, it throws the “semi-finished block” that only “Top Blockspace” has transactions to the L2 p2p network broadcast;
6. After MEV Searcher receives the “semi-finished block”, it can construct its own profitable “transaction bundle” according to the order of transactions in the “Top Blockspace”;
7. MEV Searcher sends its “transaction bundle” and “bid” to L2 Block Builder;
8. At this time, the Builder has received the “semi-finished block”, and he will put the “Trading Bundle” of the “highest bid” into the “Botton Blockspace”;
9. Finally, the Builder needs to go through the L2 Mev Boost process, and the Sequencer will accept the “highest value block” with its designated “Top Blockspace”.

! [scale70] (https://img-cdn.gateio.im/resized-social/moments-40baef27dd-430469c41e-dd1a6f-7649e1)

Summarize

By dividing “Blockspace” into two parts, user transactions can be protected in the “Top Blockspace”, and Mev Searcher can go to the “Bottom Blockspace” together, which protects user transactions from being hamful mev, and Sequencer can maximize the “block income”. However, this solution pays extra computational costs, mainly because users need to calculate a “zk proof” for their own time puzzles and Sequencer needs to solve the “time puzzles” provided by each user.

We can make a comparison with Arbitrum’s previous transaction ranking strategy that allows Mev Searcher to obtain the highest priority of 0.5s through higher bids. Compared with the scheme proposed in this paper, the Arbitrum method is characterized by:

1. Save computing resources;
2. MEV Searcher cannot see the transactions in the block (Private Mempool);
3. The user’s transaction will still be queued.

Finally, by the way: The reason for “zk proof” is to prevent the sequencer from being attacked by DDOS.