According to U.Today, a recent discussion between Polygon co-founder Mihailo Bjelic and Ethereum's Vitalik Buterin has brought Ethereum's data storage issues to the forefront. The debate began when Bjelic questioned the decision to switch Ethereum to Verkle trees, considering the likelihood of its zk-fication.
Buterin responded by stating that the current data structure, Merkle Patricia trees, with worst-case 300 MB witness sizes, is not suitable for zk-fication. These trees are the primary data structure currently used within the Ethereum blockchain. However, they have their limitations, and the plan is to replace them with Verkle trees, which offer shorter proofs and increased bandwidth.
Bjelic argued that the current structure could handle the 300 MB benchmark, based on Polygon zkEVM's capacity. He noted that the protocol could achieve this with a total block time of 12 seconds, but it would require approximately 10 server-class CPUs.
Buterin countered by explaining that Merkle trees were not designed to be compatible with ZK technology. On the other hand, Verkle trees with a vector element are designed to achieve this goal. He emphasized that switching to Verkle trees would aid in making the Layer 1 (L1) zk-friendly.
In the broader Ethereum architecture, ZK protocols are primarily used as Layer-2 scaling solutions. Despite these L2s offering better throughput, scalability, gas fees, and data storage, Ethereum still struggles with scaling up to meet demand. Buterin highlighted that converting Ethereum to a ZK-tech compatible network would only be possible using Verkle trees. If successful, Ethereum could potentially benefit from the impacts of the Dencun upgrade in the long term.