Why do some people believe that only the nodes miners run matter?

[Wind_FURY]

All full validating nodes, mining or not, validate all the transactions and blocks in the network and check if they follow the rules, correct?

Then I assume that that my node is as good as a miner's node.

But why do some people believe that only the miners' nodes matter and are very quick to cite this?



Is there a technicality in how the whitepaper is written?

Plus if the miners create the block and their nodes do the only validations that matter then what's the point?

[mu_enrico]

I have been waiting for someone to ask this question since I have the same question, and I'm just a newbie who reluctant to start a new topic.

From what I read arguments against non-mining full node, as follows:

It doesn't make the network stronger. It's actually the opposite. A non-mining full node only acts as a "witness." It doesn't have the ability to put/discard transactions in the blockchain. Thus, it only delays sending transactions to the node that really matters.

Parties who against the non-mining full node often cite Satoshi's mail "Only people trying to create new coins would need to run network nodes.  At first, most users would run network nodes, but as the network grows beyond a certain point, it would be left more and more to specialists with server farms of specialized hardware.  A server farm would only need to have one node on the network and the rest of the LAN connects with that one node."

I really hope Bitcoin guru here can give me some lesson.

Source:
https://medium.com/@olivierjanss/why-non-mining-full-nodes-are-a-terrible-idea-ad3c49f7a7b6
https://www.mail-archive.com/cryptography@metzdowd.com/msg09964.html

[Syke]

All full validating nodes, mining or not, validate all the transactions and blocks in the network and check if they follow the rules, correct?

Then I assume that that my node is as good as a miner's node.

But why do some people believe that only the miners' nodes matter and are very quick to cite this?

Yes, your node is useful, but no, it is not "as good as". The blockchain really only takes a step forward when a block is created. You cannot create blocks.

[TheQuin]

From what I read arguments against non-mining full node, as follows:

It doesn't make the network stronger. It's actually the opposite. A non-mining full node only acts as a "witness." It doesn't have the ability to put/discard transactions in the blockchain. Thus, it only delays sending transactions to the node that really matters.

There's more to it than this but the basic reason people want to run their own node is trust. Your own node can validate all transactions. If you rely on other peoples nodes then you are putting trust in them to validate transactions on your behalf.

[DooMAD]

Technically, there was no such thing as a "non-mining node" when the whitepaper was written.  I suspect the CPU -> GPU -> ASIC arms race evolved far more quickly than Satoshi envisaged.  The original idea was that everyone running a full node would be mining and everyone else would be using SPV.  But, as with seemingly all economies, we had to go and invent the "middle class".  Thus, non-mining nodes came into being.

This is why some people proposed re-writing the whitepaper to make that distinction clearer.  Non-mining nodes are important because mining began to centralise at too rapid a pace.

[ranochigo]

It doesn't make the network stronger. It's actually the opposite. A non-mining full node only acts as a "witness." It doesn't have the ability to put/discard transactions in the blockchain. Thus, it only delays sending transactions to the node that really matters.

Other than the mentioned point, nodes only accept transactions/blocks that follows the network rules and thus are essentially enforcers of it. If non-mining nodes doesn't exist, then everyone is assumed to be running a SPV client. SPV clients do not validate blocks and they assume that every transaction in a given block of the longest chain is valid. SPV clients connects to several nodes for this purpose. (Sybil) Attacks against are made more expensive with more nodes as a single honest node would make it extremely expensive. If there were no non-mining nodes at all, miners would have total control of the Blockchain. They can essentially work together and change the protocol rules of Bitcoin and its not what we want.

[ABCbits]

Let's not rely on whitepaper or Satoshi's idea since it was created long time ago under different condition. Things such as one-CPU-one-vote have too many obstacle to be realized even if Bitcoin only can be mined with CPU.

While it's true that "non-mining node" isn't as useful as as "mining node", "non-mining nodes" are required to verify many things such as :
1. Know which one is longest chain, it's really useful when chain-split occurred unintentionally. The example is Bitcoin-qt upgrade 0.8 in 2013.
2. Detect invalid transaction in block, since you can't detect it only from block header which SPV uses.
3. Know that other nodes/miners change their Bitcoin protocol intentionally.

But people should note that not every user should run full nodes, but user should be able to run full nodes if they want to without expensive/specialized hardware.

More info : https://freedom-to-tinker.com/2015/07/28/analyzing-the-2013-bitcoin-fork-centralized-decision-making-saved-the-day/

All full validating nodes, mining or not, validate all the transactions and blocks in the network and check if they follow the rules, correct?

Then I assume that that my node is as good as a miner's node.

But why do some people believe that only the miners' nodes matter and are very quick to cite this?

Yes, your node is useful, but no, it is not "as good as". The blockchain really only takes a step forward when a block is created. You cannot create blocks.

Theoretically you can create blocks, but only lack the amount of PoW needed.

[amishmanish]

This is one of the main arguments we all come across from those who feel there is no problem in any increase in block size. Moore's law being the other. The supporters of big blocks believe that there is no danger in relegating all mining and consensus power to specialized hardware manufacturers like Bitmain. The reason they cite is that miners have incentive to keep the structure intact and to not attack the chain. They assume that the miners or miner community have no reason to attack the chain completely neglecting the possibility of takeover and disruption by authoritarian governments or other political upheavals.

A non mining full node helps one to verify transactions as they accept only well-formed transactions that follow all the rules. Imagine a small merchant that decides to use a cryptocurrency for his business. Does he need to be a mining node? Not necessarily. On the other hand, does he need the ability to independently verify his transactions? Does he need to ensure that all incoming transactions follow the consensus rules as agreed by the code and there are no unknown changes made to it?? The answer to that would be yes. Both of these are things that only a full node can do. Now imagine thousands of these merchants and users running their own non-mining full nodes while a centralized only-mining entity decides to change the rules and propagate a chain with an increased reward. All those "users and merchants" reject those transactions. The changed transactions become part of the longest chain nonetheless because, well, mining nodes decide what appends on to the blockchain. But if those transactions are rejected by the merchants and users running non-mining full nodes, do they have any value, despite being part of the longest chain?? I don't think so. This is the only way to ensure that mining nodes stay in line even when they don't want to or are being forced not to.

Far too many problems in this world have happened because people are told that they should have no problem in relegating responsibility to a trusted organization. Take an example of any trusted authority in the world and you can find numerous examples where that power has been abused to suppress the common man. Tax authorities, Legislatures, Regulators in education, health and insurance, all of them have abused their power because "They are supposed to act as per the rules".

The whole point of bitcoin and being your own bank was to change this paradigm. Running a full node (mining/ non-mining) does ensure that.

[HeRetiK]

From what I read arguments against non-mining full node, as follows:

It doesn't make the network stronger. It's actually the opposite. A non-mining full node only acts as a "witness." It doesn't have the ability to put/discard transactions in the blockchain. Thus, it only delays sending transactions to the node that really matters.

[...]

I doubt that having a multitude of non-mining full nodes makes the network slower. If anything it makes the network more resilient and possibly less prone to orphaned blocks due to enabling multiple network paths to and between mining nodes.


Either way, I've looked through the medium post you are referring too, and I don't think their arguments hold quite up:

It prevents Bitcoin from scaling

Running full nodes (or the discussion about the need for full nodes) does not prevent Bitcoin from scaling. Keeping blocks small is what enables people to run full nodes, which is necessary to keep the network censorship-resistant. (Cue the whole small blocks vs big blocks discussion all over again)

Other scaling solutions exist (eg. Lightning) that do not require pruning the network down to a handful of datacenters.

It is also worth noting that the network congestion didn't last, showing that increasing the block size now would have been too early. Also the drop of Bitcoin's dominance below 50% is more likely due to the growth of the cryptocurrency market in general, rather than people fleing from Bitcoin in earnest.

They make the network insecure

Being forced to connect "directly to a trusted mining full node" is exactly where Bitcoin would become permissioned. You now have a single node that you put your trust into and if they deny your transaction your shit out of luck.

Going through "a maze of 'non-mining full nodes'" that you don't need to trust is the only way to make sure your transaction actually gets propagated across the network and thus relayed to all -- or at least most -- miners. Not just that the ones that you "trust".

I also don't see how having the sender connect directly to mining nodes would make accepting 0-conf transactions safer for the recipient, as even within a network of mining nodes only, the recipient has no guarantee that all miners have received the same transaction. Worse, still, the recipient's "trusted mining node" may not even be aware of an ongoing double spend attack.

It makes upgrades harder

It does indeed make network upgrades harder. However I'd still argue that in the case of decentralized, immutable blockchains network upgrades shouldn't come too easy -- otherwise we'd just have auto-updating nodes, which is something that Core has explicitely decided against for fear of centralization.

The question of node updates aside, the drama of recent times was more a question of politics and self-governance, rather than technology.

[Traxo]

Craig Wright[1] had documented that a new block propagates within roughly 1 second to something on the order of ~98% of the hashrate. But the 1000s of nodes in the remaining ~2% of the hashrate can see significant delays. Mining nodes prioritize their connectivity to nodes which generate block solutions, because being late on mining a new block solution costs money.

Thus non-mining nodes are essentially ignored by the consensus system of Nakamoto proof-of-work. That is why user activated fork is nonsense. The only way the non-mining users can overrule the miners is to change the proof-of-work hash of the system (which is centralization because voting and politics is centralized manipulation).

Your non-mining full node can enable you to objectively validate transactions (which is not possible in ANY proof-of-stake systems), but can't have any appreciable impact on the consensus.

[1] Yeah I know he makes mistakes and is irritating, but a few of his points are correct.

[achow101]

Craig Wright[1] had documented that a new block propagates within roughly 1 second to something on the order of ~98% of the hashrate. But the 1000s of nodes in the remaining ~2% of the hashrate can see significant delays. Mining nodes prioritize their connectivity to nodes which generate block solutions, because being late on mining a new block solution costs money.

Thus non-mining nodes are essentially ignored by the consensus system of Nakamoto proof-of-work. That is why user activated fork is nonsense. The only way the non-mining users can overrule the miners is to change the proof-of-work hash of the system (which is centralization because voting and politics is centralized manipulation).

Your non-mining full node can enable you to objectively validate transactions (which is not possible in ANY proof-of-stake systems), but can't have any appreciable impact on the consensus.

[1] Yeah I know he makes mistakes and is irritating, but a few of his points are correct.

The time to propagate blocks is completely unrelated to this. How does taking more time for a block to reach a non-mining node matter? If that non-mining node rejects the block, and if enough non-mining nodes reject the block, then the miners are still on a fork that is not being used so their coins are effectively worthless. The block propagation time is completely irrelevant; the point is not to prevent mining nodes from seeing those blocks, rather it is so that miners who are using a different blockchain from the rest of the network (which occurs regardless of block propagation times) are mining useless coins. Block propagation time has nothing to do with consensus.

[odolvlobo]

I believe it has been shown that most nodes are at most one or two hops from a miner. That and the fact that nearly all miners connect directly to each other reduce the value of non-mining nodes as part of the consensus protocol.

However, I feel that the reduced value in that role does not diminish the importance of a wallet's ability to validate the blocks and transactions it receives.

[Wind_FURY]

I have been waiting for someone to ask this question since I have the same question, and I'm just a newbie who reluctant to start a new topic.

From what I read arguments against non-mining full node, as follows:

It doesn't make the network stronger. It's actually the opposite. A non-mining full node only acts as a "witness." It doesn't have the ability to put/discard transactions in the blockchain. Thus, it only delays sending transactions to the node that really matters.

Of course it does. If a full validating node that is not a miner receives a non-valid transaction then it will not propagate it to the rest of the network.

Full validating nodes also assure censorship resistance and can give checks and balances to miners' nodes.

Parties who against the non-mining full node often cite Satoshi's mail "Only people trying to create new coins would need to run network nodes.  At first, most users would run network nodes, but as the network grows beyond a certain point, it would be left more and more to specialists with server farms of specialized hardware.  A server farm would only need to have one node on the network and the rest of the LAN connects with that one node."

Source:
https://medium.com/@olivierjanss/why-non-mining-full-nodes-are-a-terrible-idea-ad3c49f7a7b6
https://www.mail-archive.com/cryptography@metzdowd.com/msg09964.html

That is the road to centralization and that would be bad for Bitcoin.

That's my understanding.

[ir.hn]

So the deal is this.  Let's say your full node finds a flaw in bitcoin.  How would you get everyone to accept this?  Well what you do is you have to get your corrected ledger to be accepted by the network.  How do you do this?  You have to fork the current chain where it went wrong and replace it with another corrected chain, and the chain must be longer than the one you are replacing since that is the requirement the protocol sets for a ledger to be accepted.  

So in other words, you have to win blocks for your version of the correct ledger to be accepted by the network.  And to win blocks you must be mining.

You finding a flaw doesn't mean diddly in other words.  You have to put your money where your mouth is and that is by getting your ledger accepted by the network which requires winning blocks.

Also non mining full nodes make the network slower but more robust.  So it is positive in my opinion and you can also be assured you can send and receive bitcoin whenever and wherever you want.  With a SPV wallet you have to hope that other nodes even allow you to connect and sometimes they don't.

[mu_enrico]

There's more to it than this but the basic reason people want to run their own node is trust. Your own node can validate all transactions. If you rely on other peoples nodes then you are putting trust in them to validate transactions on your behalf.

<...>
If there were no non-mining nodes at all, miners would have total control of the Blockchain. They can essentially work together and change the protocol rules of Bitcoin and its not what we want.

<...>
1. Know which one is longest chain, it's really useful when chain-split occurred unintentionally. The example is Bitcoin-qt upgrade 0.8 in 2013.
2. Detect invalid transaction in block, since you can't detect it only from block header which SPV uses.
3. Know that other nodes/miners change their Bitcoin protocol intentionally.

<...>
The supporters of big blocks believe that there is no danger in relegating all mining and consensus power to specialized hardware manufacturers like Bitmain. The reason they cite is that miners have incentive to keep the structure intact and to not attack the chain. They assume that the miners or miner community have no reason to attack the chain completely neglecting the possibility of takeover and disruption by authoritarian governments or other political upheavals.
<...>

<...>
Being forced to connect "directly to a trusted mining full node" is exactly where Bitcoin would become permissioned. You now have a single node that you put your trust into and if they deny your transaction your shit out of luck.
Going through "a maze of 'non-mining full nodes'" that you don't need to trust is the only way to make sure your transaction actually gets propagated across the network and thus relayed to all -- or at least most -- miners. Not just that the ones that you "trust".
I also don't see how having the sender connect directly to mining nodes would make accepting 0-conf transactions safer for the recipient, as even within a network of mining nodes only, the recipient has no guarantee that all miners have received the same transaction. Worse, still, the recipient's "trusted mining node" may not even be aware of an ongoing double spend attack.
The question of node updates aside, the drama of recent times was more a question of politics and self-governance, rather than technology.

<...>
Full validating nodes also assure censorship resistance and can give checks and balances to miners' nodes.
<...>
That is the road to centralization and that would be bad for Bitcoin.
That's my understanding.

I think now I have a better understanding about this issue. The opposing parties of non-mining nodes have the underlying assumption (hypothetical) that miners will never cheat. Miners will never destroy their own investment, therefore users can trust mining nodes. That being said, efforts to validate transactions and detect invalid transactions in the blocks are just a waste of time since users can trust mining nodes.

Meanwhile, the supporting parties of non-mining nodes have the underlying assumption (hypothetical) that there is a probability of cheating and a possibility of takeover and disruption by authoritarian governments or other political upheavals.

I guess parties who against the non-mining nodes would have their "Oh Shit" moment when their hypothesis could be proven false.

[HeRetiK]

I think now I have a better understanding about this issue. The opposing parties of non-mining nodes have the underlying assumption (hypothetical) that miners will never cheat. Miners will never destroy their own investment, therefore users can trust mining nodes. That being said, efforts to validate transactions and detect invalid transactions in the blocks are just a waste of time since users can trust mining nodes.

[...]

It is worth noting that there's a difference between trusting mining nodes in general (ie. the network as a whole) and specific, individual nodes (ie. a directly connected trusted mining node). Being able to trust the network as a whole while not having to trust individual miners is why decentralization is so fundamental. If you have to rely on trusted individual miners you might as well stop using cryptocurrencies altogether.

[ir.hn]

mu enrico,

no that is not it at all.  Realists realize that non mining nodes are good for the node holder and network, but not by validating the network.  The point that those of us who don't idolize non mining nodes is we realize that they have no vote in the consensus of the blockchain and so if they found something wrong in the network, they would be powerless to change or fix it.  However miners do have the power to fix problems in the blockchain.

[Traxo]

The time to propagate blocks is completely unrelated to this. How does taking more time for a block to reach a non-mining node matter?

It’s the converse implication that matters and was the point being made. That is that non-mining nodes are irrelevant because they can’t insert themselves into the hashrate-weighted network to cause propagation delay between mining nodes.

The block propagation time is completely irrelevant

Incorrect. Any delay for mining nodes on receiving a new block is some lost profit. The selfish mining attack exemplifies that delayed (withheld in that case) blocks matter for profit.

the point is not to prevent mining nodes from seeing those blocks, rather it is so that miners who are using a different blockchain from the rest of the network (which occurs regardless of block propagation times) are mining useless coins.

Your statement does not make non-mining nodes relevant. Nor does it refute the point that was being made, which is that non-mining nodes are irrelevant. What can they impact? They can’t even impact propagation delay.

I believe it has been shown that most nodes are at most one or two hops from a miner. That and the fact that nearly all miners connect directly to each other reduce the value of non-mining nodes as part of the consensus protocol.

That agrees with our point.

[amishmanish]

the point is not to prevent mining nodes from seeing those blocks, rather it is so that miners who are using a different blockchain from the rest of the network (which occurs regardless of block propagation times) are mining useless coins.

Your statement does not make non-mining nodes relevant. Nor does it refute the point that was being made, which is that non-mining nodes are irrelevant. What can they impact? They can’t even impact propagation delay.

The statement does exactly that. You choose to keep ignoring the point being made on the technicality of "non-mining nodes cannot append onto the blockchain". Everyone knows that. Where do you think bitcoin derives it value from? Bitcoin does not derive its value from miners or the longest chain. It derives value from the usage and non-third party trust ensured by the PoW security that miners provide.

Non-mining full nodes are supposed to be run by 1.) Users 2.) Merchants/ Small businesses 3.) Exchanges. Those are the entities that give it value. Transactions rejected by such full nodes have no takers because they will appear invalid to the users, exchange & merchants . That is why miners will never dare mine a chain separate from that acceptable to the rest of the world, i.e. "Non-mining Full nodes".

On the other hand, if you leave only the miners to run full nodes while the rest of the world is supposed to assume they are honest, it changes everything. That is why it is stressed that you should be running a full node if you can afford to.

[Wind_FURY]

There's more to it than this but the basic reason people want to run their own node is trust. Your own node can validate all transactions. If you rely on other peoples nodes then you are putting trust in them to validate transactions on your behalf.

<...>
If there were no non-mining nodes at all, miners would have total control of the Blockchain. They can essentially work together and change the protocol rules of Bitcoin and its not what we want.

<...>
1. Know which one is longest chain, it's really useful when chain-split occurred unintentionally. The example is Bitcoin-qt upgrade 0.8 in 2013.
2. Detect invalid transaction in block, since you can't detect it only from block header which SPV uses.
3. Know that other nodes/miners change their Bitcoin protocol intentionally.

<...>
The supporters of big blocks believe that there is no danger in relegating all mining and consensus power to specialized hardware manufacturers like Bitmain. The reason they cite is that miners have incentive to keep the structure intact and to not attack the chain. They assume that the miners or miner community have no reason to attack the chain completely neglecting the possibility of takeover and disruption by authoritarian governments or other political upheavals.
<...>

<...>
Being forced to connect "directly to a trusted mining full node" is exactly where Bitcoin would become permissioned. You now have a single node that you put your trust into and if they deny your transaction your shit out of luck.
Going through "a maze of 'non-mining full nodes'" that you don't need to trust is the only way to make sure your transaction actually gets propagated across the network and thus relayed to all -- or at least most -- miners. Not just that the ones that you "trust".
I also don't see how having the sender connect directly to mining nodes would make accepting 0-conf transactions safer for the recipient, as even within a network of mining nodes only, the recipient has no guarantee that all miners have received the same transaction. Worse, still, the recipient's "trusted mining node" may not even be aware of an ongoing double spend attack.
The question of node updates aside, the drama of recent times was more a question of politics and self-governance, rather than technology.

<...>
Full validating nodes also assure censorship resistance and can give checks and balances to miners' nodes.
<...>
That is the road to centralization and that would be bad for Bitcoin.
That's my understanding.

I think now I have a better understanding about this issue. The opposing parties of non-mining nodes have the underlying assumption (hypothetical) that miners will never cheat. Miners will never destroy their own investment, therefore users can trust mining nodes.

They also discounted the reality that mining has become more centralized over the years by the mining cartel. The "one CPU one vote" as Satoshi envisioned does not apply anymore.

That being said, efforts to validate transactions and detect invalid transactions in the blocks are just a waste of time since users can trust mining nodes.

But you do not have to. You can validate them yourself by running a full node.

Meanwhile, the supporting parties of non-mining nodes have the underlying assumption (hypothetical) that there is a probability of cheating and a possibility of takeover and disruption by authoritarian governments or other political upheavals.

That is too much but yes, it would be better for the users to validate transactions and check if they follow the rules. It is also good for maintaining decentralization.

I guess parties who against the non-mining nodes would have their "Oh Shit" moment when their hypothesis could be proven false.

It is false. Anyone should have that choice and ability to validate transactions and blocks, mining node or not. Anyone who convinces you "no" should be validated. Hahaha.