Ethereum Transaction Processing Delays Measured With 79.5% Accuracy

Computing the processing time of a given transaction on the Ethereum network involves obtaining two critical timestamps: the pending timestamp, which marks when the transaction is first seen in the network, and the processed timestamp, which indicates when the transaction is included in a block. However, obtaining accurate values for these timestamps is challenging due to the decentralized nature of the Ethereum blockchain.

Each miner node in the Ethereum network maintains its own pending pool, and these pools are rarely exposed to the outside world. Even if nodes are set up to monitor the network, the peer-to-peer architecture means that transactions may take a long time to propagate, leading to inaccurate pending timestamps. For instance, Ethermine, one of the largest Ethereum mining pools, has over 300,000 nodes distributed globally, making it difficult for any single node to capture the exact moment a transaction is first seen.

To overcome these challenges, researchers relied on Etherscan, a popular Ethereum blockchain explorer, to approximate the pending timestamp. They equated the pending timestamp of a transaction with the instant at which an Etherscan node first sees it. This approach involves several steps, including the transaction being built, submitted, broadcasted, and eventually seen by an Etherscan node. The pending timestamp is then recorded from the Etherscan's Pending Transactions webpage.

However, when transactions take a long time to be processed, Etherscan updates the transaction details page to include a new field called "Time First Seen," which contains an older timestamp. Researchers used this field as the pending timestamp when it is shown, as it provides a more accurate representation of when the transaction was first seen in the network.

To evaluate the accuracy of the collected pending timestamps, researchers conducted an experiment where they submitted transactions to Ethereum using various gas price categories and recorded the timestamp at which they were sent. They then compared these submitted timestamps with the pending timestamps obtained from Etherscan. The results showed that Etherscan becomes aware of pending transactions in 1 to 2 seconds in 79.5% of the cases, with a small and stable lag between the submitted timestamp and the pending timestamp.

This study highlights the challenges and methods involved in measuring Ethereum transaction processing delays. The findings suggest that while Etherscan provides a good approximation of the pending timestamp, there is still room for improvement in accurately capturing the exact moment a transaction is first seen in the network. This information is crucial for end-users who rely on accurate transaction processing times to make informed decisions about their transactions on the Ethereum network.