EIP-4444, historical data expiry proposal, has become one of the major goals for Ethereum developers next year.
During this week's Ethereum ACDC conference, the issue of rapid historical data growth was revisited after the completion of the Cancun upgrade. Although the growth rate has decreased due to the upgrade, it is still too fast. Therefore, developers have reached a consensus to advance the importance and schedule of EIP-4444. Members of the Ethereum Foundation summarized the conference content.
This article is compiled and summarized. For any doubts, please refer to the original text.
Table of Contents
Background: Introduction to State and Historical Data
Definition
According to Paradigm's research, participants are provided with the basic concept that State and Historical Data are two different sets of data:
- State is a set of data needed to create and validate new Ethereum blocks, consisting of contract bytecode, contract storage, account balances, and account nonce.
- Historical Data is a data set needed to synchronize nodes from the genesis block to the latest block, consisting of blocks and transactions.
How State and Historical Data Grow
State and Historical Data place three pressures on the hardware of nodes:
- State Growth: Accumulation of new accounts, new contract bytecode, and new contract storage data.
- Historical Data Growth: Accumulation of new blocks and transactions.
- State Access: Relevant read and write operations when constructing and validating blocks.
Every time a transaction is executed on Ethereum, all resources used by that transaction are priced in gas fees. Therefore, Ethereum's gas fee limit restricts all forms of on-chain activities, affecting block sizes and the number of operations per block.
With more bytes per block, historical records grow faster. On the other hand, the more operations per block, the greater the state access rate, and typically, the greater the state growth rate.
How State and Historical Data Growth Affects Nodes
These pressures lead to an increasing burden on the hardware of nodes, specifically in four aspects:
- Network Read: The amount of upload and download speed that nodes must maintain for stable consensus with the network.
- Storage Size: The amount of data that nodes must store permanently for constructing, validating, and distributing blocks.
- Memory Size: The amount of cache data that nodes must keep in memory to stay synchronized with the latest blocks.
- Storage Read: The amount of read and write operations per second that nodes must perform to stay synchronized with the latest blocks.
Historical Data Makes a Comeback!
In this meeting, the issue of historical data growth was once again brought up.
Cross-chain Bridges as the Main Cause of Historical Data Growth
At the ACDC meeting, developers @notnotstorm and @gakonst analyzed the growth of Ethereum state data and historical data, indicating that the development speed of historical data is ten times faster than that of state data, mainly due to various cross-chain bridges.
It is expected that historical data will significantly affect the overall storage structure of Ethereum, leading to an increased hardware burden on nodes.
London Upgrade Slows Growth Rate
The London upgrade provides some assistance in slowing down historical data growth. It has reduced approximately 50% of the historical data growth related to cross-chain bridges, which is about 33% of the overall growth.
How does the London upgrade affect users? Key points summarized at the offline meetup Ramble Bar
However, the growth rate of Ethereum historical data is still about ten times that of state growth rate.
EIP-4444 Gains Importance
Therefore, in this meeting, most developers unanimously agreed to continue researching EIP-4444 to introduce historical data expiration features. The ideal goal is to stop providing pre-merged historical data on the Ethereum p2p layer around next year.
For EIP-4444 to be implemented, other methods for downloading historical records and standardizing storage formats are already in development.
Recommended Reading: Ethereum Releases 2024 Roadmap, Continuing Forward in the World Settlement Layer Direction
Reason for recommendation: The article mentions the positioning of EIP-4444 in the Ethereum roadmap. Originally, developers believed that the historical data growth issue could be alleviated through other means, postponing EIP-4444. After the London upgrade, historical data expiration will be a higher priority task.