Stats on compact block reconstructions

Prefilling transactions in compact block announcements has the potential to significantly improve the efficiency and predictability of transaction relay within the Bitcoin network. Effective strategies for prefilling include selecting transactions that had to be specifically requested, those sourced from an extra transaction pool, as well as transactions not present in the local mempool but which proved useful after being announced.

A recent paper explores this approach by implementing a method closely aligned with b10c's proposal. Further advancements are reflected in a draft pull request on Bitcoin Core (PR #35558), where additional considerations—such as the TCP window size—are integrated into the construction of prefilled transaction lists. This refinement aims to optimize network throughput and minimize redundant data exchange during block propagation.

The experimental evaluation detailed in section 7 of the aforementioned paper provides insightful real-world results. The study deploys twenty nodes on the Bitcoin mainnet, evenly split between sender and receiver roles. To ensure fair performance assessment, receiver nodes are exclusively connected to their corresponding sender nodes, each receiving a distinct variant of the compact block relay protocol. All participating nodes maintain direct connections to the main Bitcoin mainnet, ensuring that observed behaviors accurately reflect live network conditions.

Alternative approaches have also been investigated, such as those documented in m4ycon’s review and a separate academic publication (see here). The latter suggests prioritizing transactions in the prefilledtxn field based on higher fee rates, larger sizes, and shorter durations in the mempool. Experimental results indicate that prefilling the most recently received mempool transactions is especially effective at reducing the likelihood of requiring additional transaction requests (getblocktxn). Conversely, prioritization by transaction fee or size alone appears to offer minimal improvement in this regard.

While methodologies across these studies vary considerably, making direct comparison challenging, there is strong evidence supporting heuristics that focus on prefilling transactions previously missed in compact blocks. Such strategies can lead to notable improvements in relay efficiency without incurring unnecessary resource usage.