EIP-7691 : Blob throughput increase

Why this EIP neeeded?

• Ethereum is scaling through a rollup-centric roadmap.
• The Dencun upgrade introduced blobs (EIP-4844), and initial monitoring showed network stability with the new gas target and a decrease in re-org rates.
• The inclusion of the IDONTWANT message in Gossipsub (Ethereum's p2p messaging protocol) reduces unnecessary data transmission, saving network bandwidth.
• This EIP aims to support Ethereum's scaling until more efficient solutions are developed like peerDAS, with blob size adjustments based on network stability post-Dencun.

EIP Proposal and Functioning

• Introducing the following fields :

• At PECTRA_FORK_EPOCH, the consensus layer clients will use MAX_BLOBS_PER_BLOCK_ELECTRA and TARGET_BLOBS_PER_BLOCK_ELECTRA.
• At PECTRA_FORK_EPOCH, the execution layer will use MAX_BLOB_GAS_PER_BLOCK, TARGET_BLOB_GAS_PER_BLOCK, and BLOB_BASE_FEE_UPDATE_FRACTION_PRAGUE.
  • MAX_BLOB_GAS_PER_BLOCK and TARGET_BLOB_GAS_PER_BLOCK will replace their counterparts from EIP-4844.
  • BLOB_BASE_FEE_UPDATE_FRACTION_PRAGUE will replace BLOB_BASE_FEE_UPDATE_FRACTION from EIP-4844.

Impact on Blob Base Fee Adjustments

In each block, there are two possible scenarios:

Before EIP-7691

• The target-to-max blob ratio was 1:2.
• If the block contained full blobs (reaching max capacity), the blob base fee increased by ~12.5%.
• If there were no blobs in the block, the blob base fee decreased by ~11.1%.
• This mechanism encouraged users to store data in blobs when demand was low.

After EIP-7691

• The target-to-max blob ratio changed to 2:3.
• With full blobs, the blob base fee now increases by ~8.2% (previously ~12.5%).
• If there are no blobs, the blob base fee now decreases by ~14.5% (previously ~11.1%).
• This adjustment is driven by the introduction of specific value of BLOB_BASE_FEE_UPDATE_FRACTION_PRAGUE, which ensures a more balanced response to blob usage.

Risks & Mitigations

Risks

1. Increased network bandwidth usage, especially for validators with limited resources.
2. The fork could cause temporary network congestion.

Mitigations

1. Conduct simulations on both mainnet and testnet to test large blocks and blobs.
2. Monitor the network closely during the fork to identify and address potential issues.