Cluster Mempool: underrated improvement in Bitcoin Core?

[Antidote47k]

Most people in Bitcoin tend to focus on soft forks, new address types, or price moves. But some of the changes that actually matter long-term usually happen quietly at the policy and network level, and they don’t get much attention.

Cluster Mempool is one of those things. It’s coming into Bitcoin Core v31, and unlike a soft fork, it doesn’t touch consensus rules. It basically changes how nodes organize and evaluate unconfirmed transactions inside the mempool. From what I’ve read, the main goals are better package handling, improved fee estimation, and more efficient block building.

What stood out to me is that changes like this can seriously improve how Bitcoin feels to use, without most users ever realizing anything changed under the hood. Things like fee prediction or transaction relay don’t sound exciting, but they affect almost every user interaction with the network.

Source: https://www.bitcoingate.net/news/bitcoin-core-31-cluster-mempool-privacy-2026?

It made me start wondering:

Do we eventually get to a point where Bitcoin improvements are mostly not about consensus changes anymore, but more about these quieter infrastructure upgrades like package relay, Erlay, mempool policy changes, etc.?

Feels like the protocol might already be mature enough that the real gains now come from how efficiently everything runs rather than changing the rules themselves.

[d5000]

Do we eventually get to a point where Bitcoin improvements are mostly not about consensus changes anymore, but more about these quieter infrastructure upgrades like package relay, Erlay, mempool policy changes, etc.?

Feels like the protocol might already be mature enough that the real gains now come from how efficiently everything runs rather than changing the rules themselves.

I think I agree but with a catch. There are probably a few major consensus upgrades that are either necessary or would improve Bitcoin significantly, so I don't expect a total "ossification" of the protocol in the next years.

One example are covenants. They could make L2s like Ark and sidechains easier to deploy and very useful for mass adoption. I think the probability is high there will be at least one consensus upgrade for that purpose. On the other hand, perhaps covenant-like structures and L2s could also be implemented via a mechanism based on the current consensus rules, like BitVM. I'm unsure which path is better.

But where Bitcoin's consensus has to react always is to new threats, with the quantum threat perhaps the most well known. So in this area I expect a few more consensus upgrades, some of them probably controversial.

In general though I think your stance is reasonable. Another example is Utreexo.

[Kruw]

Cluster mempool is an improvement, but it's marginal and only applies in certain situations (unconfirmed txs with many recipients). I've been excited about it since it improves coinjoin confirmation reliability:

Child pays for parent only measures the combined fee rate of ancestors and descendants. Cluster mempool measures the combined fee rate of ancestors, descendants, AND siblings. It's strictly better for both users and miners because it sorts competing fees more accurately. This means that when multiple users spend outputs of an unconfirmed coinjoin, they will get confirmed slightly faster, and miners who upgrade to cluster mempool will create blocks that harvest slightly more in fees.

Awesome, mempool.space has already updated to graph the cluster relationships: https://mempool.space/tx/683eaa33dd0679348f4bc5d36beaae27803c464a45f2b50a7262bd717ff783f5?cpfp=advanced

[gmaxwell]

Cluster mempool is important--  but from an ecosystem perspective the importance is at least half that if the Bitcoin project didn't develop it the largest miners/pools would develop alternatives internally and increase their profits relative to others which is a centralizing force.  Open development that makes sure the BEST mining is available to all mitigates that risk.

There are many other improvements still needed to fully deliver on that improvement due to a decade of poor management of the mining functionality in Bitcoin core-- largely due to Luke-jr blockading/toxifying attempts to improve things.  E.g. cluster mempool sped up block creation enormously but now we have the fact that getblocktemplate spends all its time serializing transaction data that isn't needed for mining-- removing that alone makes it _5x faster_, and for years miners/pool ops have proposed changes to make things like that optional in GBT and luke-jr railroaded them.  The bitcoin project still seems to lack the spine to ban luke-jr from the repository, so anyone that wants to improve mining in core can expect to be harassed by luke-jr and his particular non-negotiable opinions about how exactly each and every detail of mining should work.

In any case...

There is a lot of good opportunity for non-consensus improvements but most of them will be improvements at the margin and risk mitigations that prevent bad outcomes but don't have an immediate in hand payoff.  For example, there is a lot of room to improve Bitcoin node behavior to make transaction censorship even more obviously futile--  but the payoff there is that some attacks which aren't currently happening and may not happen won't happen because they won't be possible.

Some risks in bitcoin exist in the forum of flaws in the consensus rules which can only be fully eliminated through consensus rule changes-- such as timewarp attack or fake confirmations through the 64-byte transaction's masquerading as interior leaves in the hash tree.  Fortunately there should be no political dimension to these fixes but unfortunately some people now seem to want to use any tool they can find to attack specific developers or, perhaps, Bitcoin in general.

For consensus changes that effectively add functionality-- they outright increase the utility of Bitcoin and so there is no replacement for that.  The only reason that the failure to move forward with any of these changes recently isn't a dire situation is because such a large percentage of the Bitcoin market is just coins sitting in exchanges or in ETF custody or is otherwise speculative ownership which is effectively indifferent to actually using Bitcoin, and similar is true for other cryptocurrencies largely reducing competitive threat.  I think it would be foolish in the extreme to be lulled into complacency by this.