Transactions Withholding Attack

[AnonyMint]

jajansen,

The cartel profits along the way because customers leave the non-cartel because the transactions for non-cartel customers are delayed by an ever increasing delay as the cartel's % of the hashrate rises.

The cartel profits by keeping all their transaction fees to themselves starving the network of them, thus reducing the network hashrate due to lower funding for all the other miners, thus increasing the cartels % of the mining hashrate (did you forget to read upthread that the cartel is mining?).

I hope you see I just described a spiral in favor of the cartel.

Amazon and Walmart together control a large chunk of transactions in the USA. If they jumped in early (e.g. 2016 or so) before most other merchants had, they could command a huge percentage.

[1715064435]

1715064435

[1715064435]

1715064435

[1715064435]

1715064435

[murraypaul]

The cartel profits along the way because customers leave the non-cartel because the transactions for non-cartel customers are delayed by an ever increasing delay as the cartel's % of the hashrate rises.

The cartel starts with less than 50% of the hashrate, therefore their transactions are delayed on average longer than non-cartel transactions are, therefore, by your argument, customers should leave them for the non-cartel parties.
You have handwaved at this by saying that the cartel can just accept transactions without waiting for confirmation, but that non-cartel merchants can't, without giving any good reason why.

The cartel profits by keeping all their transaction fees to themselves starving the network of them, thus reducing the network hashrate due to lower funding for all the other miners, thus increasing the cartels % of the mining hashrate (did you forget to read upthread that the cartel is mining?).

The cartel loses money because they refuse to process transactions (and thus take transaction fees) from other parties.
Other parties profit because they are (unwittingly) withholding their own transactions from the cartel.
The cartel has less than 50% of transactions, therefore they lose more fees than they withhold from the non-cartel parties.
Therefore, by your argument, the cartel mining hashrate will reduce due to lower funding.

I hope you see I just described a spiral in favor of the cartel.

Exactly the opposite.

Take away all the cartel fluff, and what you are basically saying is that someone who can fund a major portion of the network hashrate, and take constant losses due to refusing transactions, can delay transaction processing.
That isn't a new idea.
You've really failed to show any link between that, and a rational reason for Amazon et al to attempt such an attack.
The purpose of the attack would be to damage Bitcoin.
The purpose of Amazon is to make money.
Damaging Bitcoin does not make Amazon money.

[murraypaul]

Amazon and Walmart together control a large chunk of transactions in the USA. If they jumped in early (e.g. 2016 or so) before most other merchants had, they could command a huge percentage.

Which is of course why Amazon and Walmart have formed a cartel to control credit card payments, denying their fees to existing players like Visa and Mastercard, with such low payment costs that other merchants have been forced to join their cartel.
Oh wait, they haven't done that, have they?

[RodeoX]

Interesting idea OP.  

There is one disincentive to using this "attack". If Amazon used such a technique they could hurt the value of bitcoin and their own bottom line. Amazon does not want to hurt bitcoin it wants to profit from it. I suspect they will develop a bitcoin payment gateway that adds value with services such as escrow. That is where they can make a scalable percentage into perpetuity, rather than a short term mining trick that drives users away from BTC.

[jajansen]

AnonyMint,

I can go pretty much with murraypaul's answer, thats how i would say things go.

The only thing that I thought about, regarding this so called (theoretically inexistent cartel in free market) cartel, is that they could attempt to mine their own transactions, and thus, having a buffer of transactions always being able to fill their blocks 100% with transactions, and thus mining more efficently...
As I current understood, each block has a certain amount of transactions, but they are not always 100% full... The trick would then be as a mining company to always have blocks 100% filled with transactions...and this will give you an advantage...and I guess this is only possible if you have a loyal backing that gives you all transactions and just is ok with waiting more than average for the confirmation. (so that there is always a buffer of transactions such that blocks are always full)...

Anyway dont know if the above written is correct, and also dont know if it has any serious implications. In anyway, they will have more than average delayed confirmation times. (not that this confirmation time is of any importance really to amazon or the customer)
So to become the biggest miner, they should then slowly increase their business, always being sure to fill blocks more efficient than others, with the backing of the obedient transactions of amazon that will simply wait untill its their turn....

Or is all above complete nonsense?

[AnonyMint]

jajansen,

I don't think we are making progress. I will bow out of trying to reexplain what I have already explained. I understand the attack very well, but it would probably require many flowcharts and explanations to get normal people to understand it. I can carry that all in my mind, no need for the flowcharts, but most people can't see things in their mind's eye. They need illustrations. I have the advantage that I understand it because it was developed in my mind. If I were reading this thread for the first time, maybe I wouldn't get it either.

[ashaw596]

OK, I've skimmed over a bit, but I don't see the point of this argument. So we're saying that in 30 years amazon is gigantic. So the point is they're linking their payment thing with their own client with no transaction fees.

It isn't that this would not be an issue, its just that its a terribly stupid way of doing it.

The easiest way to get away from fees would just to do their transactions off network. Get people to send btc to an online "wallet" and just do everything off network once they get the btc balance. You can't stop them from doing this and the transactions are just as feeless as if they did the stupid own wallet.

The wallet idea is stupid. No one would ever use something with a 50 minute confirmation time unless somebody accepted a 0 confirmation spend, which only amazon would accept so they might as well take their transaction off network on their own wallets.

If someone sees something I don't, please respond.

Its not like you could stop them or that it would be illegal or against the way a lot of the current stuff is handled ....

[AnonyMint]

So the point is they're linking their payment thing with their own client with no transaction fees.

The easiest way to get away from fees would just to do their transactions off network.

No you missed the entire point. Try again.

Amazing to me that such an easy concept for me is so difficult for so many readers.

[ashaw596]

So the point is they're linking their payment thing with their own client with no transaction fees.

The easiest way to get away from fees would just to do their transactions off network.

No you missed the entire point. Try again.

Amazing to me that such an easy concept for me is so difficult for so many readers.

Actually I don't think I did. If they had their own client without TX. Lets say they have 20% of the network has hash rate.  It would take 50 minutes for a single confirmation. No one would use it unless the person they send to takes zero (or maybe 1 confirmation).  At that point it would only be used for Amazon purchase (assuming we're using Amazon as a stand in).  At that point they might as well just make the transactions off block. Where did my reasoning go wrong?

[Rassah]

Sorry AnonyMint, but despite you having a 160 IQ, and being very intelligent and smart, and being able to see things in your mind and conceptualize better than anyone else here, based on what I have been reading here, I can still see things even above and wider beyond you, am still more intelligent than you. MoonShadow may not have understood your premise exactly, but you are still not grasping the offliine thing completely, as well as a few other things that have been discussed before. To me, it's as if you are able to see the entire clearing up to the trees, and are proclaiming yourself to the the total expert on this clearing you have just wandered into, while I can see the trees, the forests, and the oceans beyond

So, let's see if I understand your proposed attack correctly:

We start with a normal steady state, where miners are only competing on hashing power, all transactions are broadcast from privately owned clients, everyone, including miners, can hear all transactions, and block rewards are insignificantly small, with all mining rewards coming exclusively from minning fees.

Amazon, a large corporation that may have reached the size and power of a cartel, steps in and creates their own custom AmazonWallet, which has the following features:

• People can deposit and store their own bitcoins into the AmazonWallet
• Amazon allows people who use their AmazonWallet to shop on their website, possibly even giving them a discount for using their wallet
• AmazonWallet may or may not charge fees, and may charge zero fees as incentive to get people in
• Amazon owns and operates their own mining hardware, which may, for example, account for 20% of the total hashing power
• All transactions created by AmazonWallet are sent exclusively and directly to Amazon's provate mining farm, and never broadcast to the rest of the bitcoin network until Amazon's miners manage to include it in a block
• AmazonWallet still allows you to send bitcoins to any other outside wallet that doesn't belong to Amazon, but the transaction still has to go through, and be mined by, Amazon's private mining farms
• Amazon may choose to use their 20% of world's total mining capacity to mine only AmazonWallet's transactions, keeping all other transactions out of the blocks they mine

Do I have it right so far? Anything I missed or misunderstood?

[MoonShadow]

MoonShadow may not have understood your premise exactly,

Rassah, I think you are well aware that I understood his premise at least as well as he thinks that he understood his premise.

[Rassah]

MoonShadow may not have understood your premise exactly,

Rassah, I think you are well aware that I understood his premise at least as well as he thinks that he understood his premise.

When you were arguing about the details of the AmazonWallet, it seemed you have missed that the wallets would ONLY connect to Amazon servers, and thus only broadcast all their transactions to Amazon's miners. You were arguing about how P2P works, but these wallets will obviously not be P2P.

[MoonShadow]

MoonShadow may not have understood your premise exactly,

Rassah, I think you are well aware that I understood his premise at least as well as he thinks that he understood his premise.

When you were arguing about the details of the AmazonWallet, it seemed you have missed that the wallets would ONLY connect to Amazon servers, and thus only broadcast all their transactions to Amazon's miners. You were arguing about how P2P works, but these wallets will obviously not be P2P.

I didn't miss that point, I took that as evidence that he didn't understand how the p2p network actually worked, that it was in Amazon's own vendors' interests that transactions be broadcast widely regardless of whether or not there was a profit to be gained from doing so.  Thus, it would be in the vendors' own interests to undermine the cartel's goals.  Amazon uses an exclusive payment system now, but most vendors can still do business outside of that system using their own websites.  Amazon is like a huge shopping mall, some people go there just to look around, and some small specialty shops benefit from being seen there.  That doesn't mean that they don't have stores that aren't in the mall.

[SomeWhere]

AntonyMint.

I come from a more technical background, but I do believe I understand the economic implications of the Bitcoin technology and I do also clearly see the point you are trying to make with your attack.

I do however believe that your attack is not viable in the long run...or rather, it's negligible.

There are, I think, two possible scenarios of how the cartel could withhold the transactions from the rest of the network, and a hybrid approach using both depending on whether the transaction is internal to the cartel network or not.


1) The cartel wallet would only connect to nodes that use the same wallet. The only part of this segregated cartel part of the network that would connect to the complete Bitcoin network would be the mining servers. They would be configured to farm all attractive transactions, while not forwarding any to the rest of the network, effectively keeping the transactions from their cartel nodes for themselves.

I suppose this is a rather unrealistic scenario on the technical front, as only a single ill-behaved node in the cartel section of the network sending transactions to a single member of the non-cartel part of the network would propagate most of the transactions from the cartel network onto the non-cartel network before being mined. A few of these nodes - and I don't doubt someone in the Bitcoin community would make it happen - and the networks would effectively be joined again. So let's move on.


2) I would assume that the scenario you have in mind would involve a custom cartel wallet, that directly sends the transactions to one or several cartel mining servers. At the same time, the cartel mining servers would exchange transactions. In this case, the only connection that allows the cartel part of the network to siphon revenue of the non-cartel part of the network are the mining servers of the cartel, and they are effectively the only part of the cartel network that is actually operating as Bitcoin nodes.

Effectively, this would stop the cartel wallet software from enabling transactions to the rest of the network. In this case, I would argue that the cartel would basically implement an off-chain payment system, which additionally records transactions on the blockchain with a delay. This comes down to nothing but blockchain bloat.


The real issue with your attack is, though, that, as long as any one node of the cartel network is running the Bitcoin protocol on some level, the transactions can be funneled back into the non-cartel network quite easily. As soon as the cartel tries to control its own part of the network to an extent where the nodes are no longer part of the Bitcoin network, they will de facto implement their own payment network with their own hybrid Bitcoin/cartel wallet. Personally, I would then consider the fees to be cartel fees, not Bitcoin fees. As they will also require proportionally more hashing power to verify their own transactions, their take from the Bitcoin network would not really create an imbalance in my view.

The question then is, at what point do you still consider fees attached to such a transaction as revenue of the Bitcoin network. Basically, on a basic level, what you are saying is that somebody who makes additional profit (in any way) can use those profits to increase their share of the mining power in the network. This is already possible nowadays, in much simpler ways than through your attack scenario.

Basically, your attack falls apart the moment other factors than just the revenue from mining play a role. As Bitcoin miners are likely to find other applications, such as heating, or - as mentioned by posters before - can be paid by some other company to keep mining while prioritizing their transactions, they will generate extra revenue, too, which, in turn, they can also use to increase their share of the mining power in the network. In fact, every big mining player will strive to diversify his income streams. I would even say that a lot of big mining operations will not solely be mining businesses anymore. For many, it might not even be their main activity.

In the end, the imbalance created by your attack will just play a small role in the overall economic balance of the mining market and will just be one extra revenue stream next to many others. The mining part of business might become secondary and the profits derived thereof might no longer be relevant to the economic balance of the market, which will be dominated by the same economic forces as the more traditional markets. The only consequence of such an attack would be to starve off businesses based solely on mining - something I think is a natural development even without your attack scenario.

In the end, I can come up with many technical ways to reduce the impact of your attack, but in a way you are right. I don't think it's realistic to remove this attack vector completely without modifying the core Bitcoin protocol. I do however believe it will simply not matter enough and the implications will be significantly less drastic than you imagine.

[AnonyMint]

So the point is they're linking their payment thing with their own client with no transaction fees.

The easiest way to get away from fees would just to do their transactions off network.

No you missed the entire point. Try again.

Amazing to me that such an easy concept for me is so difficult for so many readers.

Actually I don't think I did. If they had their own client without TX. Lets say they have 20% of the network has hash rate.  It would take 50 minutes for a single confirmation. No one would use it unless the person they send to takes zero (or maybe 1 confirmation).  At that point it would only be used for Amazon purchase (assuming we're using Amazon as a stand in).  At that point they might as well just make the transactions off block. Where did my reasoning go wrong?

Refuted upthread. (not intending to obfuscate nor be unhelpful, but really I can't rewrite the thread for every person)

[AnonyMint]

Sorry AnonyMint, but despite you having a 160 IQ,

I never claimed that. My IQ is some where between 125 and 150 IQ. Roughly I can conceptualize what at least 9 out of 10 people can't. It could go as high as I sometimes see what only 1 in 100,000 can.

IQ or g is not a complete model of intelligence.

So, let's see if I understand your proposed attack correctly:

We start with a normal steady state, where miners are only competing on hashing power, all transactions are broadcast from privately owned clients, everyone, including miners, can hear all transactions, and block rewards are insignificantly small, with all mining rewards coming exclusively from minning fees.

Amazon, a large corporation that may have reached the size and power of a cartel, steps in and creates their own custom AmazonWallet, which has the following features:

• People can deposit and store their own bitcoins into the AmazonWallet
• Amazon allows people who use their AmazonWallet to shop on their website, possibly even giving them a discount for using their wallet
• AmazonWallet may or may not charge fees, and may charge zero fees as incentive to get people in
• Amazon owns and operates their own mining hardware, which may, for example, account for 20% of the total hashing power
• All transactions created by AmazonWallet are sent exclusively and directly to Amazon's provate mining farm, and never broadcast to the rest of the bitcoin network until Amazon's miners manage to include it in a block
• AmazonWallet still allows you to send bitcoins to any other outside wallet that doesn't belong to Amazon, but the transaction still has to go through, and be mined by, Amazon's private mining farms
• Amazon may choose to use their 20% of world's total mining capacity to mine only AmazonWallet's transactions, keeping all other transactions out of the blocks they mine

Do I have it right so far? Anything I missed or misunderstood?

Mostly correct, except don't assume that Amazon will require customers to hold balances in Amazon wallets. I assume it would be in their interest to keep all balances on the Bitcoin block chain. Reasons were stated upthread.

And remember that Amazon can accept 0-confirmations for their transactions, since the customer is beholden, so no delay while waiting for their mining servers to win a block.

And remember that keeping the non-cartel transactions out of 20% of the blocks, means 20% of the time the non-cartel transactions will be delayed by 20 minutes, 4% of the time by 30 minutes, and 0.8% of the time by 40 minutes. Would you shop at the non-cartel if 0-confirmation was not an option with that tiny merchant and the delay was not acceptable (e.g. buying fast food at drive through, buying almost anything retail)?

Thus certain types of tiny merchants will need to join a cartel.

It is the way online (delayed shipment) cartels can capture all the retail (in person, instant delivery) market.

Also note the deleterious effect on the non-cartel miners. They have to compete with the difficulty driven 20% higher, yet they don't get that 20% of the revenue. Thus over time the non-cartel miners have less profits and thus can buy less hardware and thus the cartel continually increases that 20%. Once they exceed 50%, they control (the original fork of) Bitcoin.

This is a Tragedy of the Commons outcome, everyone must join a cartel to compete. It is form of selfish mining, although technically and economically quite different than the recently announced selfish-mining attack (which btw I showed can be easily defeated, which is a 180-degree-turn from my original position on it).

jajansen's mistake is he assumes the design of Bitcoin is free market. I am attempting to show the design is a Tragedy of the Commons.

[AnonyMint]

MoonShadow may not have understood your premise exactly,

Rassah, I think you are well aware that I understood his premise at least as well as he thinks that he understood his premise.

When you were arguing about the details of the AmazonWallet, it seemed you have missed that the wallets would ONLY connect to Amazon servers, and thus only broadcast all their transactions to Amazon's miners. You were arguing about how P2P works, but these wallets will obviously not be P2P.

I didn't miss that point, I took that as evidence that he didn't understand how the p2p network actually worked, that it was in Amazon's own vendors' interests that transactions be broadcast widely regardless of whether or not there was a profit to be gained from doing so.  Thus, it would be in the vendors' own interests to undermine the cartel's goals.  Amazon uses an exclusive payment system now, but most vendors can still do business outside of that system using their own websites.  Amazon is like a huge shopping mall, some people go there just to look around, and some small specialty shops benefit from being seen there.  That doesn't mean that they don't have stores that aren't in the mall.

And I agreed that Amazon would broadcast transaction P2P when the destination recipient was not a member of the cartel. That was not a refutation of my attack.

[AnonyMint]

And remember that keeping the non-cartel transactions out of 20% of the blocks, means 20% of the time the non-cartel transactions will be delayed by 20 minutes, 4% of the time by 30 minutes, and 0.8% of the time by 40 minutes. Would you shop at the non-cartel if 0-confirmation was not an option with that tiny merchant and the delay was not acceptable (e.g. buying fast food at drive through, buying almost anything retail)?

10 minutes is too long to wait for an in-person transaction. (Hell, 10 seconds is too long to wait for an in-person transaction.)

The attack applies even more so (because relative effect on convenience is more noticeable) if we reduce the block period to 1 minute (which is probably about the minimum that can work technically at current network propagation speed).

So 20% of time delayed by double, which means 20% longer checkout and drive thru lines.

Don't forget the other optional attack vector is the cartel offers lower transaction fees (which are spiraling higher and higher on Bitcoin), thus the customers prefer the cartel and also the non-cartel miners are cheated of income while the difficulty is higher. Cartel profits from economies-of-scale from more customers thus offsetting loss of fees.

In-person retail transactions will have to be accepted with 0 confirmations, or no one will use Bitcoin for them. There are lots of possible ways to do this. The two most likely ones are probably to treat a 0-confirmation transaction like a check which could possibly bounce, and require the purchaser to show ID, and/or to use a third party who guarantees against double-spends. The latter could be done using a split-key so the third party doesn't have unilateral access to the funds.

Another destruction of decentralization and taking us back to the system we have now.

Also note the deleterious effect on the non-cartel miners. They have to compete with the difficulty driven 20% higher, yet they don't get that 20% of the revenue.

Miners never get the revenue of the blocks they don't mine.

The cartel withholds transactions from non-cartel miners that win the next block.

[AnonyMint]

I come from a more technical background, but I do believe I understand the economic implications of the Bitcoin technology and I do also clearly see the point you are trying to make with your attack.

I do however believe that your attack is not viable in the long run...or rather, it's negligible.

There are, I think, two possible scenarios of how the cartel could withhold the transactions from the rest of the network, and a hybrid approach using both depending on whether the transaction is internal to the cartel network or not.


1) The cartel wallet would only connect to nodes that use the same wallet. The only part of this segregated cartel part of the network that would connect to the complete Bitcoin network would be the mining servers. They would be configured to farm all attractive transactions, while not forwarding any to the rest of the network, effectively keeping the transactions from their cartel nodes for themselves.

I suppose this is a rather unrealistic scenario on the technical front, as only a single ill-behaved node in the cartel section of the network sending transactions to a single member of the non-cartel part of the network would propagate most of the transactions from the cartel network onto the non-cartel network before being mined.

I did not propose the cartel will run an internal P2P network. I proposed the cartel will centralize each customer's payment from the Bitcoin chain to the cartel's balance. Then the cartel can pay cartel vendors in separate transaction(s).

2) I would assume that the scenario you have in mind would involve a custom cartel wallet, that directly sends the transactions to one or several cartel mining servers. At the same time, the cartel mining servers would exchange transactions.

The cartel centralized command and control, owns and controls the mining servers and can 100% dictate what they do.

[AnonyMint]

In-person retail transactions will have to be accepted with 0 confirmations, or no one will use Bitcoin for them. There are lots of possible ways to do this. The two most likely ones are probably to treat a 0-confirmation transaction like a check which could possibly bounce, and require the purchaser to show ID, and/or to use a third party who guarantees against double-spends. The latter could be done using a split-key so the third party doesn't have unilateral access to the funds.

Another destruction of decentralization and taking us back to the system we have now.

It's still decentralized. Any company can provide escrow services and/or no-double-spend guarantees and/or photo IDs.

That is not decentralization, because you have counter-party risk. Also big fish eat little fish, so trends towards centralization over time.

Besides retail is going away as the 3D printer rises and we will download and print. Don't think the 3D printer will be limited to simple materials and forms, they already have multi-head printers and later will surely see automated sewers and weavers, etc..

In person retail transactions will be accepted with 0 confirmations. Customers will not accept 10 minute waits or 1 minute waits. Most won't even accept 10 second waits once the novelty factor wears off. And reducing the block time is not a viable solution.

Retail is going away because you lose an hour driving there and back, and selection is limited. Instant customization is coming delivered via the internet instantly.

Anyone who thinks that Bitcoin, or any digital currency, is going to eliminate the presence of financial services companies and bring us to a world where everyone interacts with everyone else on a purely peer-to-peer basis, is living in a fantasy world.

Ha. I think you are wrong. Let's see who wins this bet.

I'm reminded, as I often am when thinking about Bitcoin, about the Internet during the early to mid 1990s, where some people thought that the fundamentally peer-to-peer nature of the Internet meant we'd never have large corporations running enormous websites.

Some people had fantasies for a long-time about PCs being ubiquitous. It didn't happen for the first 3 decades, but now with the smartphone, it is happening.

Those large sites are next on our list to slay, after we are done slaying Bitcoin.

We were basically lacking a funding model to compete with those large sites decentralized. With a better Bitcoin, we will have it. I've had many ideas about how to destroy facebook, but the barrier was always the f$cking credit card companies and Paypal bullsh8t.