Ethereum Research Report: If the execution layer retains only 1 year of active state, the state size will be reduced by approximately 78%

Foresight News reported that Ethereum execution layer researcher weiihann released a performance experiment report on State Expiry. The experiment compared the performance differences between a "full state node" and a node that "only retains one year of active state" by running approximately one year of Ethereum mainnet block load on the Geth client. The study found that reducing the state size not only lowers hardware requirements but also creates room for increasing the gas limit and throughput.

The experimental results show that if only the active state accessed within one year is retained, the database size will decrease from 359 GB to 81 GB, a reduction of about 77.5%, with the most significant reduction in the storage of Trie nodes. In terms of performance, block re-execution time was shortened by about 15%, and read performance improved significantly, especially with P50 storage read latency reduced by 46% and P99 latency reduced by 36%. In addition, tail latency also improved, with P99 block insertion time shortened by 21%, helping nodes stay in sync under load. Future research will cover comparisons with other clients, exploration of different expiry cycles (such as 6 months), and schemes that only clean up contract storage.