Centralization in mining.

[pseudospace]

Apologies if this is a topic which has already been extensively discussed, but does it not disturb anyone that in the last week ~55% of network hashrate has been shared between 2 pools? Maybe my judgement is incorrect but it seems like a disaster just waiting to happen. Is moving towards more asic resistant algorithms something which has been considering or is feesable?

It seems in the last couple of years mining has had a pretty dramatic shift into the hands of larger cooperations and governments.

Are my worries justified?

pseudospace.

[_act_]

Even till now, Foundry USA and AntPool are both having over 51% of the hash rates in total. But not that only one mining pool is having the over 51% hash rates.

It would have been good if the hash rate can be more distributed than only two pools having over 51% of mining hash rates. Let us see what will happen when Tether and Ocean that former Twitter CEO raised funds for will start operation. Tether said they are investing $500 million into bitcoin mining.

[pseudospace]

It would have been good if the hash rate can be more distributed than only two pools having over 51% of mining hash rates.

Absolutely.  Having incentives to move people away from the larger pools would be a very good start. I'm unfamiliar with what the appeal of these 2 pools are which draw such a large number of miners.

[NotFuzzyWarm]

Unlike Antpool, Foundry is *NOT* a 'public' pool that anyone can use. It is for their private mining and for Institutional miners who contract with them. The attraction for Institutional miners is simple: Dirt cheap rates and no equipment to buy. The same applies to Marathon (MARA).

Now as to popularity of Antpoo - no idea. Yes they too have Institutional contracts and with their US operations can offer them rates on-par with Foundry and MARA. I'm certain that a large part of their hashrate are Institutional customers. Question there is, why do so many much smaller miners use Antpoo? Frankly, for small miners their payout scale suxs...

[pseudospace]

Question there is why so many much smaller miners use Antpoo? Frankly, for small miners their payout scale suxs...

Yeah I'm kind of lost on what the appeal is, guessing it's simply a case of people going with the most well known option, without full knowledge of the impact their decision is having on the network. Kind of raises questions about the barrier of entry for mining being too low for those with large amounts of money to spend, but lack fundamental knowledge of how Bitcoin works.

[NotFuzzyWarm]

Kind of raises questions about the barrier of entry for mining being too low for those with large amounts of money to spend, but lack fundamental knowledge of how Bitcoin works.

And along that line the AC power needed to feed miners. eg having the skills needed to SAFELY do the wiring needed to handle it...

[pseudospace]

It's interesting to think about. I really doubt anything will change for the better (I imagine the issue will likely worsen over time) unless smaller pools are able to come up with incentives to effectively pull in miners.

This issue isn't just to limited to Bitcoin, I believe Monero has a similar issue with the majority of hashrate tied up in a minority of pools. It's unfortunate to see Bitcoin's decentralization slowly being crushed by these pools.

[FP91G]

It's interesting to think about. I really doubt anything will change for the better (I imagine the issue will likely worsen over time) unless smaller pools are able to come up with incentives to effectively pull in miners.

This issue isn't just to limited to Bitcoin, I believe Monero has a similar issue with the majority of hashrate tied up in a minority of pools. It's unfortunate to see Bitcoin's decentralization slowly being crushed by these pools.

Do you think that a private pool like Foundry will collude with other mining pools for a 51% attack? And which centralized exchange will they attack?
If you do not make transactions during the attack, then your bitcoins are safe.

[pseudospace]

Do you think that a private pool like Foundry will collude with other mining pools for a 51% attack? And which centralized exchange will they attack?

If I'm honest I don' think so. However having such a large % of the network controlled by such a small number of pools fundamentally goes against the principles of a decentralized network.

If you do not make transactions during the attack, then your bitcoins are safe.

Safe from the attack, but (unless you act quickly) not safe from the inevitable market crash that would follow news of Bitcoin undergoing a 51% attack.

[ABCbits]

Is moving towards more asic resistant algorithms something which has been considering or is feesable?

Using ASIC resistant algorithm wouldn't help since the main problem is many miner prefer to mine on popular pool or pool with big hashrate. Even Monero which use such algorithm seems to be dominated by 2 major mining pool[1].

[1] https://miningpoolstats.stream/monero

[Synchronice]

Apologies if this is a topic which has already been extensively discussed, but does it not disturb anyone that in the last week ~55% of network hashrate has been shared between 2 pools? Maybe my judgement is incorrect but it seems like a disaster just waiting to happen. Is moving towards more asic resistant algorithms something which has been considering or is feesable?

It seems in the last couple of years mining has had a pretty dramatic shift into the hands of larger cooperations and governments.

Are my worries justified?

pseudospace.

It's very bad that only commercial giants are able to mine bitcoins and also commercial giants are able to hold the biggest bitcoin reserves and own the top exchanges. I feel like Bitcoin is getting super centralized.
It's very bad that two giant pool owns more than 51% of bitcoin hashrate but it's no different from distributing it over 10 puppet pools. Also, Chinese and US pools share almost the whole market.
Chinese pools: AntPool, ViaBTC, F2Pool.
American Pools: Foundry USA, Mara Pool.

Question there is why so many much smaller miners use Antpoo? Frankly, for small miners their payout scale suxs...

Doesn't make sense, especially, if we keep in mind that these pools don't share collected transaction fees with miners and they are getting huge reward in tx fees. Doesn't really make any sense to join AntPool, they are also asking for KYC as far as I know.

[philipma1957]

It's interesting to think about. I really doubt anything will change for the better (I imagine the issue will likely worsen over time) unless smaller pools are able to come up with incentives to effectively pull in miners.

This issue isn't just to limited to Bitcoin, I believe Monero has a similar issue with the majority of hashrate tied up in a minority of pools. It's unfortunate to see Bitcoin's decentralization slowly being crushed by these pools.

Do you think that a private pool like Foundry will collude with other mining pools for a 51% attack? And which centralized exchange will they attack?
If you do not make transactions during the attack, then your bitcoins are safe.

They are attacking right now. It is fee hike attack. They do not need 51% to do it.

A 51% attack would be stupid as it would kill btc. So do a fee attack.

foundry makes 48 blocks a day.

say they deliberately make only 40 via downclock

they save power and drive fees up.

but if fees raise by 1.5 btc and they make 40 blocks that is sixty coins extra.

they lost 8 blocks at maybe 6.75 coins that is 54 coins .

a gain of six coins and huge power savings.


just wait til the ½ ing

same attack gain same 60 coins
save power
and losing 8 blocks is only 30 coins so net 30 coin profit.

[stompix]

Let us see what will happen when Tether and Ocean that former Twitter CEO raised funds for will start operation. Tether said they are investing $500 million into bitcoin mining.

At $20/th even assuming al goes into gear and not in wages, legal expenses, datacenter build, fees and many more you will get 25exas, that's 4% of the current hashrate.

They are attacking right now. It is fee hike attack. They do not need 51% to do it.

A 51% attack would be stupid as it would kill btc. So do a fee attack.
foundry makes 48 blocks a day.
say they deliberately make only 40 via downclock
they save power and drive fees up.
but if fees raise by 1.5 btc and they make 40 blocks that is sixty coins extra.
they lost 8 blocks at maybe 6.75 coins that is 54 coins .
a gain of six coins and huge power savings.

 

You're forgetting that by Foundry not mining those blocks increases the epoch time, so their other gear will mine a lot more time for the same result with diminished returns, so if there is no real power savings here, you're forcing 90% of the network gear to find the same stuff as 100% would have done normally, so you're going to have to spend more time doing it, burning more power, 15 instead of 14 days but not gaining proportionally extra blocks with the remaining online hash.

Sorry Phil, doesn't make sense to force fees up and earn more by not earning at all and spending more on power.
Besides, block data doesn't match that scenario all
https://explorer.btc.com/btc/pool/29

[starlight849]

Question there is why so many much smaller miners use Antpoo? Frankly, for small miners their payout scale suxs...

Yeah I'm kind of lost on what the appeal is, guessing it's simply a case of people going with the most well known option, without full knowledge of the impact their decision is having on the network. Kind of raises questions about the barrier of entry for mining being too low for those with large amounts of money to spend, but lack fundamental knowledge of how Bitcoin works.

Another thought, miners also like predictable earnings, and this what larger pools can offer, yes the earning share is smaller due the number of participants in the pool but the predictability is the key. If I were to invest in hardware I would need to have some reassurance of expected earnings to at least to a some degree. Larger pools are more likely to mine block than smaller pools and hence the reward is sure even if its smaller.

This is unfortunate behavior as it does not aid into decentralization. Smaller/new pools  need to issue incentives as already mentioned above. Eg. OCEAN now have zero fee for all miners. Whether it will help I'm not sure. OCEAN so far in dec 2023 found only two blocks: 819489 819242. The good thing is that there is a larger share for any participant however, as mentioned is seems unpredictable.

[philipma1957]

Let us see what will happen when Tether and Ocean that former Twitter CEO raised funds for will start operation. Tether said they are investing $500 million into bitcoin mining.

At $20/th even assuming al goes into gear and not in wages, legal expenses, datacenter build, fees and many more you will get 25exas, that's 4% of the current hashrate.

They are attacking right now. It is fee hike attack. They do not need 51% to do it.

A 51% attack would be stupid as it would kill btc. So do a fee attack.
foundry makes 48 blocks a day.
say they deliberately make only 40 via downclock
they save power and drive fees up.
but if fees raise by 1.5 btc and they make 40 blocks that is sixty coins extra.
they lost 8 blocks at maybe 6.75 coins that is 54 coins .
a gain of six coins and huge power savings.

 

You're forgetting that by Foundry not mining those blocks increases the epoch time, so their other gear will mine a lot more time for the same result with diminished returns, so if there is no real power savings here, you're forcing 90% of the network gear to find the same stuff as 100% would have done normally, so you're going to have to spend more time doing it, burning more power, 15 instead of 14 days but not gaining proportionally extra blocks with the remaining online hash.

Sorry Phil, doesn't make sense to force fees up and earn more by not earning at all and spending more on power.
Besides, block data doesn't match that scenario all
https://explorer.btc.com/btc/pool/29

it does match.  slow first 5 days fast last 9 days. end of day 5 -1%. end of 14 days +3 or 4%

[stompix]

it does match.  slow first 5 days fast last 9 days. end of day 5 -1%. end of 14 days +3 or 4%

Here are the blocks Foundry mined the last two epoch by date:

10  37    26 48
11  48    27 36
12  34    28 33
13  43    29 37   
14  35    30 31
15  35    01 49
16  44    02 43
17  43    03 45
18  43    04 53
19  50    05 30   
20  53    06 40
21  53    07 29
             08 30
             09 37

Last epoch Foundry had it's three lowest block digging days in the last 5 days, just because something happened once it doesn't make it a rule Phil.

Second, if Foundry was now back at it's full power, then it means it has run underclock forever!

Current reported hashrate is 148 Exahas, Foundry never had this hashrate ever till now, their November one day high is 140exa, so where would this extra hashrate come from? New gear and luck, the same new gear that is making them have the biggest share of the pie in months:
https://explorer.btc.com/btc/insights-pools
They are at 37.12% over the last theree days and compared to 28.24% over 1 month!

Third, don't you think it's really a bad strategy to mine more block on Friday and Saturday right before a holiday, when the reward will probably hit record lows?

[starlight849]

Feels like the Ocean Pool is slowly picking up on the speed. The hash rate for the past month almost doubled. This is cool to see in order to mix up the unwanted centralization. They have also started a new year with a new block: 823925

[philipma1957]

it does match.  slow first 5 days fast last 9 days. end of day 5 -1%. end of 14 days +3 or 4%

Here are the blocks Foundry mined the last two epoch by date:

10  37    26 48
11  48    27 36
12  34    28 33
13  43    29 37    
14  35    30 31
15  35    01 49
16  44    02 43
17  43    03 45
18  43    04 53
19  50    05 30  
20  53    06 40
21  53    07 29
             08 30
             09 37

Last epoch Foundry had it's three lowest block digging days in the last 5 days, just because something happened once it doesn't make it a rule Phil.

Second, if Foundry was now back at it's full power, then it means it has run underclock forever!

Current reported hashrate is 148 Exahas, Foundry never had this hashrate ever till now, their November one day high is 140exa, so where would this extra hashrate come from? New gear and luck, the same new gear that is making them have the biggest share of the pie in months:
https://explorer.btc.com/btc/insights-pools
They are at 37.12% over the last theree days and compared to 28.24% over 1 month!

Third, don't you think it's really a bad strategy to mine more block on Friday and Saturday right before a holiday, when the reward will probably hit record lows?

Just saying that they can do it and we can’t track it a lot. Other than daily block counts.

Besides with software that can downclock or upclock a thousand units in 1 minute. The downclock is basically doable when ever.  It also becomes cheaper when the ½ ing comes. As the penalty for losing a block reward drops from 6.25 to 3.125  is an extra incentive to manipulate fees upwards.

I think they are fine tuning it for the ½ ing. Learning how fast they can clog the mempool via under clock.

then overclock to grab the high fee blocks.

[BitcoinSoloMiner]

it does match.  slow first 5 days fast last 9 days. end of day 5 -1%. end of 14 days +3 or 4%

Here are the blocks Foundry mined the last two epoch by date:

10  37    26 48
11  48    27 36
12  34    28 33
13  43    29 37    
14  35    30 31
15  35    01 49
16  44    02 43
17  43    03 45
18  43    04 53
19  50    05 30  
20  53    06 40
21  53    07 29
             08 30
             09 37

Last epoch Foundry had it's three lowest block digging days in the last 5 days, just because something happened once it doesn't make it a rule Phil.

Second, if Foundry was now back at it's full power, then it means it has run underclock forever!

Current reported hashrate is 148 Exahas, Foundry never had this hashrate ever till now, their November one day high is 140exa, so where would this extra hashrate come from? New gear and luck, the same new gear that is making them have the biggest share of the pie in months:
https://explorer.btc.com/btc/insights-pools
They are at 37.12% over the last theree days and compared to 28.24% over 1 month!

Third, don't you think it's really a bad strategy to mine more block on Friday and Saturday right before a holiday, when the reward will probably hit record lows?

Just saying that they can do it and we can’t track it a lot. Other than daily block counts.

Besides with software that can downclock or upclock a thousand units in 1 minute. The downclock is basically doable when ever.  It also becomes cheaper when the ½ ing comes. As the penalty for losing a block reward drops from 6.25 to 3.125  is an extra incentive to manipulate fees upwards.

I think they are fine tuning it for the ½ ing. Learning how fast they can clog the mempool via under clock.

then overclock to grab the high fee blocks.

this is interesting, underclock to clog and overclock to grab the high fee, i think your on to something, that makes sense

[BlazinaBits]

I do not think that a 51% attack is likely because the pools are incentivized to keep the value of bitcoin high for their business. An attack on the network would plummet the value of Bitcoin. If this was the intent of a government entity, they would have to coordinate with other governments to also take control of large pools within their jurisdictions.

What I do see as a problem is the ability to censor transactions, or prioritize transactions by building the coming block in specific ways. This is an opaque process that does not have any accountability, and not really any consequence. There is also high chances that government entities have such transactions in mind they would like prioritized or censored.

Ocean pool is looking to solve this, I think they will be successful.