Vitalik Buterin: Ethash ASICs Not a Threat to Ethereum 🚨 This article is 7 years old...
Vitalik Buterin: Ethash ASICs Not a Threat to Ethereum
5 min read
Updated: Apr. 9, 2020 at 6:43 pm UTC
Space Image via Blockonomi
Following the release of the Bitmain E3 Ethereum ASIC miner, core developers are faced with a hard choice: Hard fork Ethereum to combat ASICs or focus development on Casper proof-of-stake which would eliminate mining.
The release of Bitmain’s E3 Ethereum ASIC has caused many in the community to voice outrage. How could Ethereum, a supposedly ASIC-resistant algorithm, fall prey to such hardware?
On April 6th, 2018, the core Ethereum developers held a meeting discussing whether it is necessary to hard fork Ethereum. According to Ethereum Improvement Proposal (EIP) 958, such a move could thwart ASIC development long enough to implement Casper.
What the Core Devs Think About the Community
According to the core Ethereum development team, the community is split into three groups regarding ASICs:
“‘(1) those who think it’s actually a security risk and they’re not entirely sure how, (2) people who think we should do nothing, and (3) people who are very emotional and say “down with Bitmain and down with ASICs.” I don’t know where that [sentiment] comes from, but I think it’s a crossover from the Bitcoin community.”‘
Overview of Ethereum Improvement Proposal 958
The high-level picture of EIP 958 is this:
“[whether] we should do something to break them [the ASICs] or whether we should sit back and just keep focusing on the things that will permanently fix this.”
The developers say that the extent of the ASIC threat is not clear. The majority of developers on the team voiced concerns that they don’t have enough information on the matter. Forking Ethereum to change the protocol is “a shot-in-the-dark” solution to the problem and may not hinder current ASICs.
Most developers on the team think that energy is best focused on preparing Casper for launch.
Vitalik’s Thoughts on Ethereum ASICs
According to Vitalik, the new ASICs aren’t that big of a threat. These ASICs are only a marginal improvement upon existing processors:
“If you look at the E3 that was released a few days ago the efficiency gains are relatively small compared to existing GPUs. My Chinese sources indicate a 220 MH/s miner costs about $2500, while Bitmain is offering a 180 MH/s for $800, which is only a 2.5x factor of improvement.”
He further speculated on how Bitmain, and others, are improving upon current GPU designs:
“[The ASICs are likely] a regular computer with a better piece of RAM and everything non-essential stripped out.”
Moreover, Vitalik made it clear that he built Ethash correctly. The underlying Ethash algorithm was built to withstand ASIC optimization, as said in his statement:
“…We did an analysis and we showed that the [Ethash] algorithm succeeds at being I.O. [input-output] bound rather than CPU [central processing unit] bound.”
Therefore, by design, building Ethash ASICs is relatively difficult. The difficulty of such development limits the proliferation and acts as an upper bound on ASIC efficiency, says Vitalik.
Vitalik Says There’s Nothing Ethereum Can Do
As mentioned above, the improvements that ASICs make in processing power are limited to Ethereum. Even if a change was made to the Ethash algorithm, current ASICs might prove already optimized for IO hard proof-of-work algorithms,
“in which case, the immediate concern is basically that, there’s nothing we can do [about ASIC resistance]… Even if we just replace Ethash entirely with Equihash or something, or whatever the hell Monero uses, ASICs would still be able to do it [and work on the network].”
Thus, even if Ethereum moved to Zcash‘s Equihash or Monero‘s CryptoNightV7, a supposedly ASIC-resistant algorithm, that would not stop the use of ASICs on the network.
Vitalik’s Short-Term Solution
Although the solutions appear limited, Vitalik did propose one possible solution, to move Ethereum to an SHA-3 based algorithm, the latest member of the Secure Hash Algorithm family released by National Institute of Standards and Technology (NIST):
“One solution I brewed up is switching to SHA-3, but with a change that drastic I’d be inclined not to. If we switch to something like that, it’s less ASIC resistant. That means that there will be ASICs for it again in 6-12 months so we’re only buying a bit of time. And then, the actual development effort of making good GPU miners for SHA-3 and getting everyone upgraded is likely to be very chaotic and detract from more important things.”
Vitalik further tempered his commitment to preparing an SHA-3 version of the Ethereum protocol in a subsequent comment. Vitalik said he is not willing to develop the SHA-3 Ethereum algorithm if it comes at the expense of other optimizations, such as better denial-of-service resistance, or Casper.
51% Attacks are Not a Threat to Ethereum
Not only that, but Vitalik also believes that a 51% attack on Ethereum is not a pressing threat:
“This is not Bitcoin; the miners are not in control here. If there comes a day where Bitmain has majority hash power and tries to use it for evil, then we will speed up Casper development and to hell with any remaining bugs. We can try really hard to release it within a week and mining rewards go down by 90%.”
As such, Vitalik claims, it is in a miner’s best interest not to attack the network.
Expediting Casper in Response to an Attack
However, expediting Casper may also introduce additional bugs. The development team said that expediting Casper comes at the expense of auditing the protocol:
“And it [Casper] is in the process of getting formally verified and checked by multiple groups of academics, but if 51% attack concerns demand it, we can totally skip all of the auditing and verification stages.”
Conditions to Hard Fork Ethereum
Another developer asked if there are a set of conditions that would warrant a response, rather than the current course of action: doing nothing. According to Vitalik, there are no indicators that the team could use see to pre-emptively hard fork Ethereum, other than a 51% attack:
“One thing to keep in mind is that ASIC manufacturers have an incentive to downplay the amount of impact they have over the network. They know the more influence they have over the network, the more political pressure we’ll have to hard fork them [the ASIC miners] away. It seems extremely likely that even if Bitmain, F2 pool, the North Korean government, or some people in Iceland or whatever, even if they have 51% of the network, us finding out before they have 51% control is quite low. So, because of that, I don’t think we are going to see flags that are redder then actually seeing an ASIC that people can buy, and anything less red than an actual 51% attack happening.”
Thus, according to Vitalik, there is nothing short of a 51% attack on the Ethereum blockchain that would indicate that the development team should implement a response.
Another core developer disagreed, and said there are two possible indicators that the team could use:
“(1) The sale of ASICs that are an order of magnitude better than what is currently available. (2) Seeing a sudden uptick in one particular mining pool suspected of ASIC mining.”
Overall Assessment of the Developer’s Thoughts
Overall, the devs do not think that current ASICs warrant expediting anything ahead of schedule; nor do they think that the current situation requires an emergency hard fork. Such moves, according to the core developers, are risky and would set-back development of a long-term fix like Casper.
That said, statements from the team also indicate that there is no consensus as to when Casper will release. Discussion during the meeting suggests it could take between six months and several years.
If you disagree with the decision, the core developers encourage the Ethereum community to conduct a carbon vote, start an EIP, or take to the forums and Reddit. With enough support, the devs said that:
“If the community wants this to happen and has a good enough reason, we can definitely do that [and hard fork Ethereum].”
In conclusion, as of right now, the core team isn’t going to do anything about Ethereum ASICs.
