I run several private mining pools, so I’m pretty familiar with how mining works.
If Flypool decided to attack Zcash, miners would very quickly realize it and would leave the pool. The company behind Flypool, which also operates a large Etheruem pool, would lose millions in revenue. If an ASIC miner decides to attack the network, the only way to stop the attack is a hard fork.The key difference is that miners choose to mine on a certain pool and can just as easily choose to leave it. If an ASIC miner acts up, there is no way that the community can easily choose to ignore their blocks (except for a hard fork).
@zooko has not been transparent in explaining his resistance to a very simple modification of the Equihash parameters. We know that such such a tweak is easy to implement and is effective against ASICs. Much smaller Equihash coins have already implemented this change with great success.
Please re-read what I wrote. I didn’t say it was likely to happen, just that it was no different than a solo miner. Solo miners could have millions in hardware investment alongside the coins that they are mining and holding, which is an equally large disincentive to not attack the network.
We know that such such a tweak is easy to implement and is effective against ASICs
A simple parameter change would not be effective at all against ASICs for more than four months, which is roughly how quickly Innosilicon created their ASIC. Bitmain could probably adapt just as quickly. A change that lasts four months isn’t worth it.
Further, it incentivizes ASIC manufacturers to not sell to the public, but to instead mine ASICs secretly so that the general public isn’t alerted to the issue. That would be the worst of all cases. Recent admissions by Innosilicon made it clear that they mined Monero in secret for months.
I run several private mining pools
ASICs allows some solo miners to skip using a pool, which might be your conflict of interest. I actually think being able to avoid pool mining is a large benefit of ASICs. It’s wrong that small miners have to pay 1-2% + latency losses to a middleman just to mine.
Flypool has 46,000 miners, while a solo pool has one miner.
If Flypool were to attack the Zcash network, the act would tarnish Bitfly’s name. They would likely lose out on tens of millions of dollars of future revenue and the company itself would lose much of its value.
This is completely different from a solo miner who has no reputation to uphold. If he attacks the network and is able to profit, he doesn’t have to worry about his name being sullied.
So, no, you are wrong. A large community mining pool and a solo miner have very different incentives. Now, you could argue that an ASIC manufacturer’s interests are aligned in the same way as a large mining pool’s. In some ways, I agree, but it’s still much harder to defend against an attacker with overpowering hardware than an attacker who has hashing power solely through community trust.
A simple parameter change would not be effective at all against ASICs for more than four months, which is roughly how quickly Innosilicon created their ASIC. Bitmain could probably adapt just as quickly. A change that lasts four months isn’t worth it.
Further, it incentivizes ASIC manufacturers to not sell to the public, but to instead mine ASICs secretly so that the general public isn’t alerted to the issue. That would be the worst of all cases. Recent admissions by Innosilicon made it clear that they mined Monero in secret for months.
So you admit that an Equihash tweak would get rid of the Bitmain and Innosilcion ASICs. So much for the whole “we don’t know how the ASICs work” argument.
I disagree with your claims:
A fork would make ASIC manufacturers more hesitant to make new ASICs.
Higher memory requirements would make ASICs less efficient relative to GPUs.
Stealth ASIC mining cannot exist in large magnitudes because the effects can be seen by looking at difficulty.
The cryptocurrency community can keep tweaking the proof of work algorithm. The cost and time required to tweak the proof of work algorithm are less than the cost and time required to make new ASICs. This asymmetry will likely result in the ASIC manufacturers giving up.
Interesting, so several times you have accused me that me with 1x (!!!) Z9 mini is biased in this Asic topic and whatever not and you just run several mining pools? And as a bonus private ones?
Just my very personal opinion and in general. Mining pools are in my opinion the most useless middle man thing needed in whole crypto. Not only do they generate a theoretical chance for abuse, scams, whatever not, but are total useless if there was a direct network which hopefully we will see with some POS variants soon…
Higher memory requirements would make future ASICs much less powerful relative to GPUs. The ASIC manufacturers and their shills are bluffing when they claim that a proof of work tweak won’t work. Unfortunately, a lot of gullible people are falling for it.
I’ll keep it simple: solo ASIC miners have invested in zec-specific hardware at significant cost. Mining pools haven’t. This has nothing to do with reputation. We all know that small miners aren’t focused on benevolent pools, they are focused on mining wherever brings them the most profit. Period.
So you admit that an Equihash tweak would get rid of the Bitmain and Innosilcion ASICs
Not long-term, nope. It would only do so for a few months at best, and then make those ASIC manufacturers less likely to announce them and sell them publicly. In this way, forking would make mining less-decentralized because smaller miners potentially wouldn’t have access to the new ASICs being developed.
A fork would make ASIC manufacturers more hesitant to make new ASICs.
Not true. There is a certain coin price risk assessment that makes ASIC manufacturing attempts inevitable, even after forks. I bet ASIC manufacturers are building new ASICs for Monero right now - and nobody will know until 6-10 months after. They can also just make more general ASICs that adapt to simple parameter changes like the ones you propose with minimal performance loss.
Higher memory requirements would make ASICs less efficient relative to GPUs.
This is impossible because GPUs are ASICs, just less efficient ones. The definition of ASIC in the context of mining is a hashing board that is more efficient than an ASIC GPU.
Stealth ASIC mining cannot exist in large magnitudes because the effects can be seen by looking at difficulty.
“in large magnitudes”? Or not at all? Because they can exist as a high percentage of the hashpower without anyone knowing; Innosilicon’s Monero miners being the best example of this. So, once again, you’re factually incorrect.
The cryptocurrency community can keep tweaking the proof of work algorithm. The cost and time required to tweak the proof of work algorithm are less than the cost and time required to make new ASICs. This asymmetry will likely result in the ASIC manufacturers giving up.
Worrying about the hashing algorithm would be a complete and ongoing waste of developer time for an outcome that would leave the Zcash network less secure.
The Zcash company and foundation have other priorities like making Zcash faster, increasing adoption, better UX/UI and improving and expanding privacy usage. Working on these fundamentals is what will increase the price of zec, not worrying about a few gpu miners who are upset that their hardware is becoming obsolete (and I’m one of them, so I’ll have to deal with it too).
The fact is that ASICs are a continual evolution of mining: cpu → gpu → asic → more future security. It’s a continual race that strives for more hashes while requiring less electricity.
I’ll keep it simple: solo ASIC miners have invested in zec-specific hardware at significant cost. Mining pools haven’t. This has nothing to do with reputation. We all know that small miners aren’t focused on benevolent pools, they are focused on mining wherever brings them the most profit. Period.
ASIC miners do not have a right to mine Zcash just because they made ASICs for Equihash. ASIC manufacturers invested in building ASIC miners for a cryptocurrency that claimed to support ASIC resistance and claimed that it would fork away from ASICs if they were developed.
Not long-term, nope. It would only do so for a few months at best, and then make those ASIC manufacturers less likely to announce them and sell them publicly. In this way, forking would make mining less-decentralized because smaller miners potentially wouldn’t have access to the new ASICs being developed.
Not true. There is a certain coin price risk assessment that makes ASIC manufacturing attempts
inevitable, even after forks. I bet ASIC manufacturers are building new ASICs for Monero right now - and nobody will know until 6-10 months after. They can also just make more general ASICs that adapt to simple parameter changes like the ones you propose with minimal performance loss.
Here’s the one thing we can guarantee with an Equihash tweak: Current ASICs will not be able to mine the new algorithm.
That’s the only guarantee.
Whether new ASICs come is uncertain. It’s likely that they won’t be able to achieve the same efficiency advantages. Either way, not being guaranteed that a fork will ward off ASICs forever is not a reason to avoid one.
This is impossible because GPUs are ASICs, just less efficient ones. The definition of ASIC in the context of mining is a hashing board that is more efficient than an ASIC GPU.
I was unclear, sorry. I meant that the efficiency advantages of ASICs over GPUs will be much smaller with greater memory requirements. The Ethereum ASICs are a good example of what Equihash ASICs would look like with different parameters.
“in large magnitudes”? Or not at all? Because they can exist as a high percentage of the hashpower without anyone knowing; Innosilicon’s Monero miners being the best example of this. So, once again, you’re factually incorrect.
Again:
The possibility that a proof of work tweak will fail to get rid of ASICs forever is not a good reason to avoid one.
Worrying about the hashing algorithm would be a complete and ongoing waste of developer time for an outcome that would leave the Zcash network less secure.
The Zcash company and foundation have other priorities like making Zcash faster, increasing adoption, better UX/UI and improving and expanding privacy usage. Working on these fundamentals is what will increase the price of zec, not worrying about a few gpu miners who are upset that their hardware is becoming obsolete (and I’m one of them, so I’ll have to deal with it too).
The fact is that ASICs are a continual evolution of mining: cpu → gpu → asic → more future security. It’s a continual race that strives for more hashes while requiring less electricity.
Mining is one of the most important things about a cryptocurreny, since it determines the distribution. If the coins go to more people, then it’s likely that the cryptocurrency will become a global currency.
Second, right now, almost all of the activity on Zcash is mining. If you don’t believe go look at a block explorer. As a corollary, hardly anyone who is not a miner is actually using Zcash. The Zcash community right now is the mining community.
Third, the Zcash Foundation receives roughly 250,000 dollars per day from the founder’s reward. They can afford to either spend a few hours working on a proof of work tweak or find a volunteer to do it (like they did for the Windows wallet). Again, this is pretty simple task.
Apparently a decision/commitment to change PoW would have to happen before Sapling deployment if it was to make it into the next update, IIRC thats slated for March.
@ChileBob I’m generally interested in an algorithm switch if it could be asserted that ASIC performance could not be significantly higher than GPU/CPU performance.
The other big issue though is that I believe any algorithm Zcash uses should require that miners commit to mining Zcash only. This is a big reason I think we should go away from GPUs entire and switch to a non-Equihash ASIC-mined algorithm that forces miners to commit. GPUs leave the network weak when it’s less profitable.
Edit: I just want to say that network security isn’t a big problem right now, but I believe it could be in 5-10 years when Zcash’s privacy features could come under attack.
If it was somehow split evenly regardless of hashing rates, then that even split gpu reward would divide amongst less gpu miners, increasing payout size
It would DE-incentive large asic farms (as far as their overhead cost goes) as their max reward is capped
Its sounds good in theory, but that’s all it is currently, data is required
For me its all about preventing too much power from accumulating in too few hands, how that’s done doesn’t really matter. I find the huge performance difference between a Z9 & 1080 disturbing.
Miners chase pennies because that’s what their game is. Algo/coin auto-switching is a great tool for them and I agree with you that its a ‘bad thing’. Maybe what we need is ‘nicehash resistance’ (only kidding).
Thinking it through a bit more - an ASIC with external RAM would be cheaper to produce than a GPU as it wouldn’t have the graphics crap, plus it would be optimised for whatever algo. We’d end up with something faster/cheaper than a high end card but not be able to completely crush GPU mining.
I agree that accumulation of power is a big problem, whether it be a mining pool with too much hashrate or a solo miner trying to take over the network at scale.
One future solution, Proof of Stake, would remove the economies of scale advantage of larger miners but it would also increase the founders reward past the expiration date (they will need funding from somewhere though).
That’s exactly where things go totally wrong and into the wrong direction in my opinion if we talk about POW mining (no matter cpu/gpu/asic).
Because we miners are just a handfull guys that got mostly into crypto because we had/have some PC skills, work on the net, can handle rigs, can put together PC’s and such stuff.
The average joe guy/girl isn’t going to mine neither with a CPU, nor with a GPU, not with a rig and not with an Asic. All of these need some skills, some more, some less.
Damn, it’s eve sometimes hard for me to find a mistake in the setup, an error, make the setup good and safe, install drivers on a rig or reverse an unwanted Windows update, nothing to explain that much, we all know the hazzles that come with mining.
And here is the problem, just because WE manage to get it somehow handled doesn’t mean the next door girl should and could as well. It’s not going to happen and therefore the distribution and spreading is totally flawed and only in favor of a handfull miners far away from being good distributed around the world.
Not even talking about the uneven starting positions when it comes to electricity costs, hardware investements many can’t afford, and and and and … [endless list here].
This might be fair distribution to you, but it’s not, no way it is. It’s totally flawed, full of mistakes, full of unfair and dangerous traps and most important, it’s a brake for wider adoption.
Seriously, the right path and new direction will be POS, maybe a lot altered, modified, combined, tweaked, whatever and i’am pretty sure vitalik came to the same conclusion and with Ethereum the first big one switching to POS it will begin a new area in crypto space and that’s absolutly the best way to be able to have it wider spread and not only in a nearly closed miner/trader/speculator community.
Enough the wallets get a bit more easier to handle, wallets for cell phones, automatic staking, payouts even for the small stakers so they get intrigued and everybody and his grandmother can be part of the crypto community with ease and without any computer tech degree.
This said, it’s best in my opinion to begin research POS instead of wasting time, resources, dev hours on asic resistance or not as it really doesn’t matter because POW is doomed to fail, expect Bitcoin eventually for which it was invented. The foundation and ZcashCo better use all resources to investigate for a longer future plan with POS than just some short-time resource consuming asic/anti-asic endless forking battle…
Vitalik Buterin:
POW IS EXTREMELY EXPENSIVE, AND FURTHERMORE IS FUNDAMENTALLY VULNERABLE TO 51% SPAWN CAMPING ATTACKS WITH NO EFFECTIVE STRATEGY FOR RECOVERING FROM ONE. SELFISH MINING IS PROFITABLE STARTING AT 25-33% HASHPOWER, AND 51% CENSORSHIP ATTACKS ARE DEFINITELY PROFITABLE.
for me the situation is as following, I`m a private guy and I don’t want to get rich, but I want to make some money, for power and others. I want to use a ASIC miner to host it effectivly outside my home.
If zChash is changing their algo, my invest is lost and I will stop mining.
Vitalik right now targets at 1500 ETH per stake and hopes that BLS ans STARKs would decrease it to 32 ETH in future. Do you want to buy at least 32 ETH to get 2 percent per year? Well you can lose all your profits from everyday volatility or earn 10+ times more than annual PoS mining profit overnight just by 20% price increase. ASIC mining is profitable only for ASIC manufacturers(even if they sell some before), PoS is not profitable for anyone. If you think profits are not important: why someone should hold ZEC now if there would be no profit in future, he can just buy it a few minutes before using it to send someone money privately and then receiver would sell it a few minutes later. But it doesn’t contribute to ZCash’s ecosystem. Today’s GPU miners are much better sollution for profit distribution than on-exchange speculators that will be only profit-gainers in PoS-based ecosystem and ASIC manufacturers in ASIC-based.
