"Preserving the advantages of compact block relay is often cited as one reason for trying to keep mempool and relay policy consistent between nodes—the more the mempools of nodes differ, the less effective compact blocks will be at minimizing bandwidth and latency."
Knots makes it hard to keep mempools similar by inconsistently refusing to forward transactions. Makes efficient block propagation harder and stale blocks more frequent, costing miners money with no benefit to the netwrk, reducing security and lowering the value of Bitcoin
The premise that some marginal improvement on miner efficiency is a higher priority than nodes being sovereign on their private mempools is... funny.
attacking Bitcoin can be done as a self sovereign act.
that doesn't mean there aren't social or economic consequences.
Also, this is a problem for miners, not node runners.
are you familiar with "noisy neighbor problem" ?
Marginal improvement on miner efficiency should be a priority to whoever cares about mining decentralization.
in the days before Gavin's character assassination, I remember liking one of his posts. it explained tech stuff really well.
Then people were all, "rah rah politics", so I began caring more about character studies and less about explaining MimbleWimble to the cashier at Scaling.
The result? Nobody invites me to waste time waiting for pizza anymore. I think it's a win-win+win, possibly win-win+win+win+win-win...-win[1]?
dumber than descartes, louder than popescu, and some old man has ALS again. good thing he's not in my rolodex. ↩
what is the actual benefit of compact block relay?
So large mining operations can have high data bandwidth connections to each other, thereby having more time to work on a block than a smaller miner that may be part of a smaller pool say running off a gas generator in South Africa or helping sustain the energy demands vs energy supply inefficiencies of some power grid somewhere remote with much slower internet connectivity.
That direct connection vs node to node gossip to eventually some small miner can mean whole minutes of time that the bigger data center style miners have working on a block that those smaller miners are still hashing on the old block.
BUT when the trade off is preventing sovereign node runners from setting their own mempool policy, it doesn't seem like individuals are going to like that regardless of the downstream effects
sus
talk about smoothing, not sustension. econometricians are smoothbrain walkandtalks, and "sats" is short for "satoshis", not the other source of dust.
Satellite internet will do more for small miners in this area I believe
dust is still the correct metaphor; noisy neighbor because whizz bang crash, lots of dust again.
does anyone feel like tacos?
Sure it can, and giving small miners the knowledge that they should add major miners as direct peers helps. We also have by default 8 two way peers and 2 one way peers. Those 2 one way peers are meant to be the high data transfer peers to help blocks get through the network faster.
However, if you're one of those smaller miners, those 2 one way peers don't exactly help the block that you found get to those bigger miners faster which again puts the smaller miners blocks at higher risk of getting orphaned.
deleted by author
How about Mara with Slipstream? Do they get a pass?
The key point here is that compact block relay is only as effective as the similarity of transaction sets across participating nodes. When mempools diverge too much the protocol loses its main advantage which is the ability to send small messages and rely on peers to already have most of the transactions. If a node implementation chooses to selectively withhold or inconsistently forward transactions this increases variance between mempools which in turn forces more full transaction data to be sent during block propagation. The result is higher latency and increased bandwidth use and in the competitive environment of Bitcoin mining this translates directly into a higher risk of stale blocks.