Is Bitcoin infrastructure too Chinese? What should be done technically?

[Wind_FURY]

Good luck in finding a solution.

Solution is almost ready, implementation is the phase.

Sincerely good luck. But I believe the idea has already failed before it has started. The miners will never signal their readiness for a hard fork that will kill them. The community would also not want to risk a chain-split.

There exists no such thing as a miner in bitcoin network, I suppose by miners you mean pools, there will be no signaling neither any negotiations, pools are the main target and they are to be extinguished.

I mean that whole part of Bitcoin. Miners, pools, ASIC makers.

How?
I admit it is one of the most sophisticated problems in bitcoin right now, getting rid of pools and transition to a truly decentralized heterogeneous network but you always find it more convenient to face a problem when both it is recognized by most people as a serious problem and there is hope/technology to overcome it.

I think both ASICs and pools have triggered enough hatred and protests again themselves and the ugly face of centralized mining is nothing to be covered by any means, the fact that they've been tolerated that long is a result of the lack of an ultimate solution especially for pools and this is what we should  focus on: preparing a solution. The community will decide how and when to adopt it.

To be quite honest, I believe there is no way for any POW coin to be ASIC-resistant. If the coin has value, there will be ASIC made.

2) Sure there may be problems with centralized pools and ASIC manufacturers,  but the money supply and transactions can still all be public despite these problems.  In other words, we are still worlds above a privately issued currency system (fiat).  We aren't talking about abuses such as issuing arbitrary amounts of monetary units for ourselves here, so it really isn't fare to compare this kind of centralization with the kind witnessed in fiat systems.  

Theoretically if there is enough of Bitcoin's "transaction processors" centralized geographically in one country, that country's government can threaten them, and make the "processors" censor some transactions that that country's government does not like.

True, and reason for concern.  However:

 1)  This shouldn't be compared to the problem of private issuance, or fiat money

But it will make the project a failure and worthless.

2)  Users already have tools to evade this kind of censorship, such as alternative chains, coinjoins, pseudonymity, stealth addresses, ...  

I believe you misunderstood. "Bitcoin's transaction processors" are the miners.

[1713493404]

1713493404

[1713493404]

1713493404

[1713493404]

1713493404

[aliashraf]

To be quite honest, I believe there is no way for any POW coin to be ASIC-resistant. If the coin has value, there will be ASIC made.

No. A memory hard algorithm is not breakable by ASICs if it is devised properly. It is always possible to have a device more specialized than commodity devices but not that far to be an order of magnitude more efficient. Ethash for instance is not cracked by ASICs despite what Bitmain says, E3 is just like 2x better than a comparable gpu rig and still the race is not over, better gpus are ready to beat E3 in next few months. Plus, commodity devices, being multipurpose, have an inherent advantage against specialized hardware like E3, this compensates for their relative inefficiency.

Another sophisticated approach to ASIC resistance is ProgPow, it is a really 'badass' algorithm that theoretically makes it almost impossible for a specialized device to have gains above 1.2x over commodity gpu and it is not enough incentive for an attacker to invest on such a device.

[aliashraf]

I still think that PoCW will need to be deployed successfully in an altcoin and tested in the wild for a good while before the bitcoin community will seriously consider it.

There are a few folks that take the pool problem seriously and that value real mining decentralization, but I think not enough to make a difference at this time.    I'm happy to be wrong on this point of course, but if others thought like I do then only a few innovative coins would have any value right now, and PoS and tokens wouldn't be a thing...

If you get a library of working code and decide to launch an altcoin with it, let me know.  I might be able to help out.

Wow! We got a contributor. Appreciate your support mate, will be in touch soon.  

As of altcoin vs bitcoin thing:
Well, ... I'm not ok with such a hypothetical dilemma. Sooner or later, we need to do something about Chinese and centralized mining, and it would be a hard-fork, undisputable. Now no matter how persuasive you are and what level of consensus is reached, there always are actors who continue mining on legacy chain, ignorantly or intentionally.

I don't know who invented this term, 'altcoin', first but as a software engineer/developer I approach bitcoin like any other system, it needs overhaul and versioning at least every few years, there is and there will be no eternal system, protocol, contract, constitution, ... ever. I understand bitcoin is a social compromise and is to be treated cautiously but it doesn't change everything and we are late already.

Now, what is bitcoin? Our decent upgraded version which is maintained directly by tens of thousands of solo miners or the legacy version that is maintained by a handful of Chinese pools?

I don't care about labels, I'm sure that getting rid of pools is the most exciting event in cryptocurrency since 2009.

[cellard]

The main problem is we would need a hardfork to implement any of these changes, which ultimately are changes on the fundamentals on bitcoin, so expect a ton of opposition.

We may need to be under actual attack for people to wake up and align against attackers with a ton of hashrate. Sadly this is how humans seem to work, we may need to be pushed into the cliff first to actually make it happen, however it's not a waste of time to have the plan B prepared.

As far as China does, as long as the cheap labor and cheap electricity is mostly there, then things will stay there I guess. But the Bitmain monopoly is alread showing big cracks, specially after being stuck with massive BCash bags.

[funkenstein]

True, and reason for concern.  However:

 1)  This shouldn't be compared to the problem of private issuance, or fiat money

But it will make the project a failure and worthless.

Do you really think so?  The project will still be censorship resistant (just not perfectly so) and have a public verifiable money supply (so no hidden counterfeiting or issuance malfeasance).  Just because there's a chance that it isn't perfect for one use case doesn't mean that it's worthless.  A helmet won't always save your life but it isn't considered worthless to all motorcyclists.   

[Wind_FURY]

To be quite honest, I believe there is no way for any POW coin to be ASIC-resistant. If the coin has value, there will be ASIC made.

No. A memory hard algorithm is not breakable by ASICs if it is devised properly. It is always possible to have a device more specialized than commodity devices but not that far to be an order of magnitude more efficient. Ethash for instance is not cracked by ASICs despite what Bitmain says, E3 is just like 2x better than a comparable gpu rig and still the race is not over, better gpus are ready to beat E3 in next few months. Plus, commodity devices, being multipurpose, have an inherent advantage against specialized hardware like E3, this compensates for their relative inefficiency.

Be careful in saying this and believing it is the "truth". The same were said of Scrypt, Equihash, and Cryptonight. But they ended to have ASICs built for them. How can you give assurance of ASIC-resistance?

I believe "ASIC-resistance" has become a buzzword.

Another sophisticated approach to ASIC resistance is ProgPow, it is a really 'badass' algorithm that theoretically makes it almost impossible for a specialized device to have gains above 1.2x over commodity gpu and it is not enough incentive for an attacker to invest on such a device.

I will give the benefit of the doubt. But I believe Jihan Wu won't.

[aliashraf]

Be careful in saying this and believing it is the "truth". The same were said of Scrypt, Equihash, and Cryptonight. But they ended to have ASICs built for them. How can you give assurance of ASIC-resistance?

I believe "ASIC-resistance" has become a buzzword.

...

I will give the benefit of the doubt. But I believe Jihan Wu won't.

The beauty of computer science is that it is no magic, just pure mathematics. No matter how expert or wealthy a cracker is, he just can't build a specialized device with much more than 2x efficiency for Ethash or 1.2x for ProgPoW. This is how it works in the real world. You take Bitmain too serious, imo.

[Wind_FURY]

Be careful in saying this and believing it is the "truth". The same were said of Scrypt, Equihash, and Cryptonight. But they ended to have ASICs built for them. How can you give assurance of ASIC-resistance?

I believe "ASIC-resistance" has become a buzzword.

...

I will give the benefit of the doubt. But I believe Jihan Wu won't.

The beauty of computer science is that it is no magic, just pure mathematics. No matter how expert or wealthy a cracker is, he just can't build a specialized device with much more than 2x efficiency for Ethash or 1.2x for ProgPoW. This is how it works in the real world. You take Bitmain too serious, imo.

No, I take what's in front of me as evidence that "ASIC-resistance" is becoming another buzzword marketing tool for altcoins that use POW.

Plus can you give us the sources on where you got "2x efficiency for Ethash or 1.2x for ProgPoW"? Thanks.

No hash power cut-off. Imagine we implant a secondary ASIC-resistant cpu/gpu friendly algo (like ProgPow) and define two parallel difficulty adjustment systems which regulate blocks mined by sha2 at 11 blocks per every 2 hours ratio while gpus are set to generate 1 block in the same window also suppose we use a gradual re-adjustment policy by introducing a decrease/increase strategy for above parameters. See? no cut-off!

Back to this. How can the network be assured that the hardware behind the hash power of the new algorithim will make up for all the ASIC? Won't buying all the generic machines needed to mine make it more expensive to maintain the same hash power?

We should show this topic to someone from the mining subforum, and make a more productive discussion out of this.

[aliashraf]

The beauty of computer science is that it is no magic, just pure mathematics. No matter how expert or wealthy a cracker is, he just can't build a specialized device with much more than 2x efficiency for Ethash or 1.2x for ProgPoW. This is how it works in the real world. You take Bitmain too serious, imo.

No, I take what's in front of me as evidence that "ASIC-resistance" is becoming another buzzword marketing tool for altcoins that use POW.

Plus can you give us the sources on where you got "2x efficiency for Ethash or 1.2x for ProgPoW"? Thanks.

ASIC-resistance has been a design principle for Pow variants for a few years and you just can't find a single algorithm being developed to be intentionally
'ASIC-friendly', hence it is not a buzzword.
On the contrary what is more journalistic and hype is the fallacy about "every algo could be ASIC-ed". It was not a fallacy if it was just about pure mathematical computations (like shA256) but it is not and there exist bandwidth/storage/memory/... bound algorithms.


As of 2x maximum efficiency for Ethash, besides checking Bitmain specification for E3 and comparing it with a 6xAMD-RX570 rig  you are welcome to check:
This reference is an Ethereum Improvement Proposal for implementing ProgPow that briefly discusses the extent of ASIC vulnerability of a handful of algorithms, including Ethash and  ProgPow.

No hash power cut-off. Imagine we implant a secondary ASIC-resistant cpu/gpu friendly algo (like ProgPow) and define two parallel difficulty adjustment systems which regulate blocks mined by sha2 at 11 blocks per every 2 hours ratio while gpus are set to generate 1 block in the same window also suppose we use a gradual re-adjustment policy by introducing a decrease/increase strategy for above parameters. See? no cut-off!

Back to this. How can the network be assured that the hardware behind the hash power of the new algorithim will make up for all the ASIC? Won't buying all the generic machines needed to mine make it more expensive to maintain the same hash power?

We should show this topic to someone from the mining subforum, and make a more productive discussion out of this.

Well, my idea is implementing an orthogonal PoW system with zero interference with legacy sha2. Both legacy and the new system coexist with no clue about each other, while each node takes care of both algos simultaneously.
1- In the beginning, ASIC miners experience 1/11 increase in difficulty such that the average blocktime increases from current 10 to 11 minutes while GPU miners experience a network with 110 minutes block time.
2- Users experience an average of 11 (10 ASIC  + 1 gpu mined)  blocks per 110 minutes, so it is the same 10 minutes block time they are used to.
4- Every 13000 blocks (like 3 months) the block time is re-adjusted for each system such that ASICs lose one block in each 110 minutes and gpus win one i.e. the ratio changes to 9:2 then 8:3 then 7:4 and so on.
5- Re-adjustment happens 9 times and stops in 1:9 ratio in favor of gpus.

As of your interest in "maintaining the same hash power", it is very important to note that security and essentially the value of bitcoin is not dependent on the network total hash rate unlike what people say carelessly, it is about the total amount of resources consumed mainly labour (human resources) and electricity/energy consumption. For instance a jump in ASIC manufacturing process won't make bitcoin more secure!

Suppose in a hypothetical scenario, all miners upgrade their hardware to new technology while each decide to maintain exactly the same hashrate they already have with the old devices, total network hashrate remains the same but energy consumption will drop significantly, because new technology consumes less resources(for asics, less power mainly).

Interestingly, because of its higher human resource consumption, gpu mining is more environment friendly than ASICs.

[Wind_FURY]

The beauty of computer science is that it is no magic, just pure mathematics. No matter how expert or wealthy a cracker is, he just can't build a specialized device with much more than 2x efficiency for Ethash or 1.2x for ProgPoW. This is how it works in the real world. You take Bitmain too serious, imo.

No, I take what's in front of me as evidence that "ASIC-resistance" is becoming another buzzword marketing tool for altcoins that use POW.

Plus can you give us the sources on where you got "2x efficiency for Ethash or 1.2x for ProgPoW"? Thanks.

ASIC-resistance has been a design principle for Pow variants for a few years and you just can't find a single algorithm being developed to be intentionally
'ASIC-friendly', hence it is not a buzzword.
On the contrary what is more journalistic and hype is the fallacy about "every algo could be ASIC-ed". It was not a fallacy if it was just about pure mathematical computations (like shA256) but it is not and there exist bandwidth/storage/memory/... bound algorithms.

But the evidence has showed that if a cryptocurrency is successful, then mining will become more efficient, and more specialized through the invention of ASIC.

Coins that make the claim that they are "ASIC-resistant" has become a buzzword.

As of 2x maximum efficiency for Ethash, besides checking Bitmain specification for E3 and comparing it with a 6xAMD-RX570 rig  you are welcome to check:
This reference is an Ethereum Improvement Proposal for implementing ProgPow that briefly discusses the extent of ASIC vulnerability of a handful of algorithms, including Ethash and  ProgPow.

Would it be wrong to believe that Bitmain and some other ASIC maker are doing their research to make ASICs for ProgPow?

No hash power cut-off. Imagine we implant a secondary ASIC-resistant cpu/gpu friendly algo (like ProgPow) and define two parallel difficulty adjustment systems which regulate blocks mined by sha2 at 11 blocks per every 2 hours ratio while gpus are set to generate 1 block in the same window also suppose we use a gradual re-adjustment policy by introducing a decrease/increase strategy for above parameters. See? no cut-off!

Back to this. How can the network be assured that the hardware behind the hash power of the new algorithim will make up for all the ASIC? Won't buying all the generic machines needed to mine make it more expensive to maintain the same hash power?

We should show this topic to someone from the mining subforum, and make a more productive discussion out of this.

Well, my idea is implementing an orthogonal PoW system with zero interference with legacy sha2. Both legacy and the new system coexist with no clue about each other, while each node takes care of both algos simultaneously.
1- In the beginning, ASIC miners experience 1/11 increase in difficulty such that the average blocktime increases from current 10 to 11 minutes while GPU miners experience a network with 110 minutes block time.
2- Users experience an average of 11 (10 ASIC  + 1 gpu mined)  blocks per 110 minutes, so it is the same 10 minutes block time they are used to.
4- Every 13000 blocks (like 3 months) the block time is re-adjusted for each system such that ASICs lose one block in each 110 minutes and gpus win one i.e. the ratio changes to 9:2 then 8:3 then 7:4 and so on.
5- Re-adjustment happens 9 times and stops in 1:9 ratio in favor of gpus.

Good luck to this idea. I hope a miner replies here.

As of your interest in "maintaining the same hash power", it is very important to note that security and essentially the value of bitcoin is not dependent on the network total hash rate unlike what people say carelessly, it is about the total amount of resources consumed mainly labour (human resources) and electricity/energy consumption. For instance a jump in ASIC manufacturing process won't make bitcoin more secure!

What?

Suppose in a hypothetical scenario, all miners upgrade their hardware to new technology while each decide to maintain exactly the same hashrate they already have with the old devices, total network hashrate remains the same but energy consumption will drop significantly, because new technology consumes less resources(for asics, less power mainly).

Interestingly, because of its higher human resource consumption, gpu mining is more environment friendly than ASICs.

What?

[aliashraf]

Would it be wrong to believe that Bitmain and some other ASIC maker are doing their research to make ASICs for ProgPow?

Yes, it would be absolutely wrong. It is not rational to make a specialized device f0r 10-20 percents more efficiency. It won't ROI.

As of your interest in "maintaining the same hash power", it is very important to note that security and essentially the value of bitcoin is not dependent on the network total hash rate unlike what people say carelessly, it is about the total amount of resources consumed mainly labour (human resources) and electricity/energy consumption. For instance a jump in ASIC manufacturing process won't make bitcoin more secure!

What?

I suppose by  'what?' You mean 'How is that?'  

Once an adversary decides to do anything bad against a network he doesn't ask about technical details, he asks about the costs, yes?

What makes bitcoin secure is not an index, say the required hashrate, it is the costs to achieve that hashrate. Suppose we got these two coins ACoin and CCoin, checking their network status we find out that ACoin has a total hashrate of 1 penta hash while CCoin is secured by like 100 terahash power, now which coin is easier to attack? We don't know! They enjoy different algorithms and we need to do a research about the costs of hiring the required power to attack each coin, don't we?

Most authors, realizing the above fact, measure the security of a network correctly by the costs of renting the required hashpower for a 50%+1 attack (for instance from Nicehash) and what I'm saying is about this cost, essentially determined by  electricity and human resource consumption.

[Wind_FURY]

Would it be wrong to believe that Bitmain and some other ASIC maker are doing their research to make ASICs for ProgPow?

Yes, it would be absolutely wrong. It is not rational to make a specialized device f0r 10-20 percents more efficiency. It won't ROI.

As of your interest in "maintaining the same hash power", it is very important to note that security and essentially the value of bitcoin is not dependent on the network total hash rate unlike what people say carelessly, it is about the total amount of resources consumed mainly labour (human resources) and electricity/energy consumption. For instance a jump in ASIC manufacturing process won't make bitcoin more secure!

What?

I suppose by  'what?' You mean 'How is that?'  

Yes of course. But out of confusion.

Once an adversary decides to do anything bad against a network he doesn't ask about technical details, he asks about the costs, yes?

A smart adversary would plan everything all the way to the end, including the costs.

What makes bitcoin secure is not an index, say the required hashrate, it is the costs to achieve that hashrate. Suppose we got these two coins ACoin and CCoin, checking their network status we find out that ACoin has a total hashrate of 1 penta hash while CCoin is secured by like 100 terahash power, now which coin is easier to attack? We don't know! They enjoy different algorithms and we need to do a research about the costs of hiring the required power to attack each coin, don't we?

Most authors, realizing the above fact, measure the security of a network correctly by the costs of renting the required hashpower for a 50%+1 attack (for instance from Nicehash) and what I'm saying is about this cost, essentially determined by  electricity and human resource consumption.

But in mining the network honestly, miners would also find the most efficient way to mine a coin. ASICs are the most cost-effective way to mine and we will always go back to the debate, "if a cryptocurrency is successful, ASICs will be developed".

Let's go back to this topic in due time. But the evidence today shows that ASICs are always made for the most successful coins. Plus they always get better.