Measuring minrelaytxfee across the Bitcoin network

Utilizing a modified version of virtu’s p2p-crawler, an analysis was conducted to observe the FEEFILTER messages from nodes that accept inbound connections. This study, carried out through scans on the 10th and 15th of September, revealed that a dominant majority of nodes advertise a minimum relay fee of 1000 satoshis. Despite this consistency, there was a notable portion of nodes, especially those within IPv4, that did not send a FEEFILTER message, hinting at operational peculiarities or the presence of spy nodes as posited by Peter Todd. The research also underscored the existence of nodes with abnormally high minimum relay fees, specifically pinpointing a value of 9170997 satoshis, raising questions about their configuration and purpose.

Further examination into the nodes not sending FEEFILTER messages unveiled a pattern suggesting potential clustering within subnets, possibly indicating a deliberate arrangement for specific network functions. This discovery was complemented by insights into nodes operating on software versions predating 2016, lacking the capability to send FEEFILTER messages due to the feature's introduction post-2016. Additionally, the possibility of nodes intentionally disabling the feefilter option for various reasons, including privacy concerns or experimental purposes, was highlighted.

The anomaly involving nodes advertising a fee filter of 9170997 satoshis led to a deep dive into Bitcoin Core's handling of minfeefilter values. The investigation revealed that this specific value results from an intentional adjustment by Bitcoin Core to prevent node fingerprinting, utilizing a mechanism known as FeeFilterRounder. This mechanism aims to optimize feerate filters by categorizing them into predefined bins, with the peculiar value of 9170997 satoshis fitting within these parameters. This behavior indicates an intricate balance between privacy, security, and efficiency in the network’s protocol design.

A significant revelation from the study was the identification of nodes running recent versions of the Bitcoin software yet failing to reply with a FEEFILTER message. This observation contradicted initial assumptions about technological constraints tied to software versions. A closer look at these nodes revealed many were in the Initial Block Download (IBD) phase during connection attempts, affecting their ability to process or respond to feefilter requests. This finding underscores the complexity of node behavior and the impact of operational states like IBD on protocol interactions within the Bitcoin network.

The evolution of node behavior concerning minimum relay fees was further documented in a subsequent network crawl conducted on December 17, 2025, showing a significant shift towards acceptance of lower feefilters, particularly among i2p and onion nodes. This trend suggests a dynamic change in network practices and configurations over time. The detailed network-by-network breakdown provided insights into the distribution of minimum relay fees across different network types, illustrating the diverse landscape of node configurations within the Bitcoin ecosystem. This comprehensive analysis, inspired by community discussions and collaborations, highlights the ongoing curiosity and efforts to understand the intricacies of blockchain networks.