Bitcoin's ASICBOOST Problem Explained [pdf]
1–10 of 131 posts
Re: Bitcoin's ASICBOOST Problem Explained [pdf]
#2Wasn't the entire point of picking SHA-256 to both check the integrity of the existing ledger (why it's a hash) AND as a proof of work lottery?
The "problem" being that instead of being a proof of actual work, shortcuts in the work were found.
It seems then that a good solution would be to force many different types of proof of work. Different basis of hashes are obvious, but maybe some type of actual work every so often as well.
Re: Bitcoin's ASICBOOST Problem Explained [pdf]
#3Larbr | Pukhtunkhwa Social Network https://www.larbr.com/
Re: Bitcoin's ASICBOOST Problem Explained [pdf]
#4Wasn't the entire point of picking SHA-256 to both check the integrity of the existing ledger (why it's a hash) AND as a proof of work lottery? The "problem" being that instead of being a proof of actual work, shortcuts in the work were found. It seems then that a good solution would be to force many different types of proof of work. Different basis of hashes are obvious, but maybe some type of actual work every so o…
The problem with the proof of work is that the machinery became dependent on the selection of transactions in the block. A "standard" miner does not have this dependency. It's not a problem (depending on who you ask) that an optimization was found, the amount of work expended is still the same.
It's not clear that switching proof of works would eliminate this class of concern. It's also clear that eliminating this class of problem is not in the interest of the miners, hence the opposition to SegWit (which does exactly that).
Re: Bitcoin's ASICBOOST Problem Explained [pdf]
#5Wasn't the entire point of picking SHA-256 to both check the integrity of the existing ledger (why it's a hash) AND as a proof of work lottery? The "problem" being that instead of being a proof of actual work, shortcuts in the work were found. It seems then that a good solution would be to force many different types of proof of work. Different basis of hashes are obvious, but maybe some type of actual work every so o…
> The "problem" being that instead of being a proof of actual work, shortcuts in the work were found.
The same "problem" occurred when people invented GPU mining, then FPGA mining, then ASIC mining. As long as SHA256 isn't broken, it doesn't matter what kind of speedup miners manage to get; whenever someone finds a big advantage, everyone else uses it soon enough.
> Different basis of hashes are obvious
"Use more hash functions" is sort of a reflexive response you get a lot from non-cryptographers, but the intuitive reason this is a bad idea is that if your hash function isn't broken, your cryptosystem shouldn't be either. It's better to fix the cryptosystem than to half-assedly patch it over by obfuscating any problems with more hash functions.
> but maybe some type of actual work every so often as well.
What do you mean by "actual work"? In all likelihood, whatever you're thinking of has none of the properties that make PoW a viable mechanism for sybil resistance.
Re: Bitcoin's ASICBOOST Problem Explained [pdf]
#6Wasn't the entire point of picking SHA-256 to both check the integrity of the existing ledger (why it's a hash) AND as a proof of work lottery? The "problem" being that instead of being a proof of actual work, shortcuts in the work were found. It seems then that a good solution would be to force many different types of proof of work. Different basis of hashes are obvious, but maybe some type of actual work every so o…
> The "problem" being that instead of being a proof of actual work, shortcuts in the work were found. The same "problem" occurred when people invented GPU mining, then FPGA mining, then ASIC mining. As long as SHA256 isn't broken, it doesn't matter what kind of speedup miners manage to get; whenever someone finds a big advantage, everyone else uses it soon enough. > Different basis of hashes are obvious "Use more has…
> everyone else uses it soon enough.
Unfortunately in this case, ASICBOOST is both patented, and adds a fair bit of complexity to mining hardware. Both these problems have the undesirable effect of increasing the barriers to entry for new asic manufacturers - an undesirable thing from the point of view of Bitcoin users.
You're quite correct that shortcuts are irrelevant, but the reality of manufacturing makes the playing field a lot less level than we'd like.
Re: Bitcoin's ASICBOOST Problem Explained [pdf]
#7Earlier quoted context omitted.
> The "problem" being that instead of being a proof of actual work, shortcuts in the work were found. The same "problem" occurred when people invented GPU mining, then FPGA mining, then ASIC mining. As long as SHA256 isn't broken, it doesn't matter what kind of speedup miners manage to get; whenever someone finds a big advantage, everyone else uses it soon enough. > Different basis of hashes are obvious "Use more has…
> everyone else uses it soon enough. Unfortunately in this case, ASICBOOST is both patented, and adds a fair bit of complexity to mining hardware. Both these problems have the undesirable effect of increasing the barriers to entry for new asic manufacturers - an undesirable thing from the point of view of Bitcoin users. You're quite correct that shortcuts are irrelevant, but the reality of manufacturing makes the pla…
Exactly, I'd go further and say that if a particular technique is patented and cannot be matched by a related technique, giving a permanent unfair advantage, it is up to the community to change the proof of work and work around the advantage imposed by a state monopoly.
Re: Bitcoin's ASICBOOST Problem Explained [pdf]
#8Earlier quoted context omitted.
> The "problem" being that instead of being a proof of actual work, shortcuts in the work were found. The same "problem" occurred when people invented GPU mining, then FPGA mining, then ASIC mining. As long as SHA256 isn't broken, it doesn't matter what kind of speedup miners manage to get; whenever someone finds a big advantage, everyone else uses it soon enough. > Different basis of hashes are obvious "Use more has…
> everyone else uses it soon enough. Unfortunately in this case, ASICBOOST is both patented, and adds a fair bit of complexity to mining hardware. Both these problems have the undesirable effect of increasing the barriers to entry for new asic manufacturers - an undesirable thing from the point of view of Bitcoin users. You're quite correct that shortcuts are irrelevant, but the reality of manufacturing makes the pla…
> both patented and adds complexity
just like any CPU or GPU solutions. just because they are more readily available you can ignore it for those cases?
Re: Bitcoin's ASICBOOST Problem Explained [pdf]
#9Earlier quoted context omitted.
> everyone else uses it soon enough. Unfortunately in this case, ASICBOOST is both patented, and adds a fair bit of complexity to mining hardware. Both these problems have the undesirable effect of increasing the barriers to entry for new asic manufacturers - an undesirable thing from the point of view of Bitcoin users. You're quite correct that shortcuts are irrelevant, but the reality of manufacturing makes the pla…
Exactly, I'd go further and say that if a particular technique is patented and cannot be matched by a related technique, giving a permanent unfair advantage, it is up to the community to change the proof of work and work around the advantage imposed by a state monopoly.
Then Bitcoin has to hard fork every time someone comes up with a new patentable idea. I don't see why people think mining must be fair. It only has to not become completely centralized.
Re: Bitcoin's ASICBOOST Problem Explained [pdf]
#10This is excellent. I've been trying to find a technical explanation of ASCIBOOST for some time. Most of the articles and discussions gloss over the details.
It is possible that this is part of the reason some miners are blocking SegWit but there is more to it. Core (the main Bitcoin development team) promised to provide a block size scaling solution and then reneged. This pissed a lot of people off and led to the blocking of SegWit. I don't see ASCIBOOST as a real problem for Bitcoin.