The biggest contributor to resisting the pressure of inscriptions?On the improvement of zkSync by the Boojum upgrade

Written by: Haotian

Recently, @zksync completed the Boojum upgrade, and it was on this premise that zkSync withstood the stress test of Operation SYNC Inscription. However, Boojum is undervalued by the market.

What performance improvements does Boojum bring? Can the criticized Decentralized Finance stability problem be solved? Next, let’s talk about my understanding:

1. The Boojum upgrade, simply understood, will allow zkSync to complete the transition from SNARK to STARK proof. The workflow is roughly as follows:

When a batch is encapsulated, these transactions are broken down into multiple specific circuits, which are then processed in parallel and high-speed to generate a large number of STARKs, which are finally aggregated into a STARK proof. Finally, the STARK proof is encapsulated into an SNARK proof and submitted to the Mainnet for verification.

This mix of STARKs and SNARKs ensures efficient processing of large volumes of transactions, while dropping the size of the data submitted to the Mainnet (SNARKs are simpler) and more compatible with the Mainnet.

The use of two proofs at the same time means that the advanced compression technology, hardware acceleration technology, Algorithm optimization, batch processing aggregation efficiency, memory and storage optimization of the Prover system will have significant performance improvements.

2. According to the @0xtaetaehoho tweet, the average data volume per transaction was 211 bytes before the Boojum update, and it can be reduced to about 68 bytes after the upgrade, and the improvement of compression technology will directly greatly increase the transaction volume of each batch in layer 2, which will greatly increase the TPS (about 450), and the gas cost of a single transaction will be drop (about 65%).

The principle is not difficult to understand, layer 2 submits proof of status data to the Mainnet calldata, due to the limited storage data on the Mainnet, layer 2’s STARK parallel processing capabilities and SNARK proof compression processing technology determine the transaction volume and gas level that can be processed by a single batch;

3. Previously, ZK-Rollup had instability problems in processing low-frequency Decentralized Finance transactions, and its native tendency was not conducive to Decentralized Finance stability. For example, the variable price of Decentralized Finance requires multiple Oracle price feeds, and if the two transactions are not batched into the same state, the resulting transaction wear will increase.

Now the transaction volume of a single batch in layer2 has increased significantly, and more Oracle data status updates can be accommodated within the block. Decentralized Finance stability issues will also be effectively solved.

As the zkSync official @anthonykrose stated, no matter how many Oracle Machine updates are contained in a Block, the entire Block state can be processed and recorded as a whole, and only need to pay the cost of one state write. This is beneficial for the low fees, high efficiency, and stability of Decentralized Finance applications on the ZK-Rollup chain.

It stands to reason that Boojum’s upgrade should be regarded as a milestone for zkSync.

On the one hand, it verifies the inference that the larger the transaction volume of the ZK system, the lower the gas fee, the better the experience, and on the other hand, it also proves that the efficient application and performance improvement of computing resources such as compression technology and hardware acceleration of the off-chain Prover system will bring infinite imagination to the ZK system.

After the Cancun upgrade EthereumMainnet blob Block Size was expected to drop the cost of layer2 batch transactions, and now the technical optimization of the ZK system itself has brought the ZK series rollup and the OP series rollup to the same level.

The point is, ZK-Rollup is much more “active” than OP-Rollup. The advantages of ZK-Rollup technology, which have always been told, have been fully proven after the Boojum upgrade.

Reference: For ZK hardware acceleration, Computing Power optimization, etc., please refer to the following research reports: