How to find nodes forwarding transactions with less than minRelayTxFee?

Question

Technically, it is up to each node operator to set the minRelayTxFee as he pleases.

Assume a transaction with less than 1Sat/vB in transaction fees and a fairly empty mempool.

Is it possible to identify nodes on the network that are likely to forward or have previously forwarded transactions with less than default minimum fees?

If a subnet of such nodes existed, miners should be incentivized to also connect to it -- at least in times of empty mempool.

EDIT:

Originally, I did not intend to discuss my assumptions about miners' incentives. However, I find that discussion rather interesting, and will happily clarify my views.

I understand that close to zero fee transactions may (currently) not incentivize miners enough. On the other hand, we can be certain that in principle miners are willing to include transactions at a fee of 999 Sat/vB. That being said, I think it is a fair assumption that a market price could be established somewhere in between, i.e. 0 < fee < 1000 Sat/vB. If that is true, transaction fees currently do NOT represent a market price, due to minRelayTxFee = 1000.

Furthermore the incentive to include non-standard transactions is likely to depend on the fiat price of Bitcoin, as the labour necessary for a transition is likely to be priced in fiat. I conclude that the incentive will grow in the future.

Finally, I do not agree with the notion that the demand for transaction is flat (static, fixed). I am absolutely certain of a demand for cheaper transactions.

EDIT:

Regardless of the transactions miners receive, nobody can force them to include those transactions in a block anyhow. If miners regard transactions as too cheap, they may very well omit them. It may not be wise to do so, because the fees, as little as they may be, will be picked up by another miner. Miners (i.e. some miners/a miner/somebody) will eventually change their parameters if the volume of non-standard transactions becomes large enough.

My entire argument is about extending the fee market to the lower side.

Answer 1

Since 0.13.0 nodes will tell their peers via the feefilter p2p message what their current minimum acceptable feerate is. Given that it's fairly uncommon for nodes to run with something other than the default, I don't think nodes use this information for preferential peering. I guess you could parse the log to identify nodes that announce something lower than the default and then add them as peers via addnode.

When I last checked in January 2022, it was obvious that no miners were mining transactions that pay less than the default minRelayTxFee. That seems rational to me: a full block of 1 s/vB transactions brings a total of ₿0.01 in fees, and transactions below the minRelayTxFee even less than that. Compared the current subsidy of ₿6.25, that's less than a sixth of a percent. My understanding is that mining pools are mostly large logistics endeavors, and they tend to have a limited contingent of programmers on staff. Given a limited pool of skilled labor, you'd probably want to prioritize other things such as energy efficiency of your operation, reducing latency, stable power supply, redundant internet connections, improving pool worker management, or any of a dozen other things that could cause your very expensive hardware to run at less than optimal performance.

Additionally, having a minimum price for blockspace is in the interest of node operators, because every transaction has the externality of needing to be transmitted, validated, and stored by every node operator. Maintaining a minimum fee rate on their mempool is also a safeguard against a malicious party wasting lots of bandwidth with unconfirmed transactions that they later replace or don't get confirmed otherwise.

There is some discussion of those findings on Twitter here: https://twitter.com/murchandamus/status/1480553169239355393

Answer 2

If a subnet of such nodes existed, miners should be incentivized to also connect to it -- at least in times of empty mempool.

Miners will never be incentivised for it. If users know there is a probability of transaction getting confirmed with fee rate below 1 sat/vbyte, many who use 1 sat/vbyte right now will start using lower fee rates. Hence, it will reduce miners revenue.