New EIP being submitted as mitigation to risk associated with EIP-1559 and potential cliff of hashrate that could drop off Eth Network. Link for EIP soon.
59 Likes
Ethereum has established a model of “minimum viable issuance” with broad community consensus. Most projections do not indicate that an increase of the block reward is required to maintain secure operation of the network- even with the successful implementation of EIP-1559.
Further, accommodation of this form only delays the inevitable as we move swiftly towards a merge of eth1 and eth2, where PoW mining rewards will be zero. IMO, we should avoid measures which encourage further investment in mining hardware by miners given this reality.
For these reasons, I don’t feel this proposal is necessary for sustaining secure operation of the network, nor does it gainfully improve the network.
29 Likes
It reads like a subsidy of users to miners to delay impending change. Why? How are miners entitled to stability and ease of change? Why start from 3 ETH, why not start from 2 ETH instead, if the purpose is simply to ease change?
There is no such thing as risk free business. Miners all over the world razed through GPU graphic cards to make profit, and that implies risk.
Their “worthless” hardware is going to easily be absorbed by the market. Even if it wouldn’t, the personal, business decisions of miners are not the concern of Ethereum users. You don’t socialize profits, then you don’t socialize losses.
23 Likes
I agree absolutely with everything that DCinvestor has said. A block reward increase is unnecessary for security, it undermines the “minimum viable issuance” monetary policy, and it’s a distraction from the Ethereum roadmap.
Accepting this proposal would establish a precedent that Ethereum’s monetary policy is more or less arbitrary and that with sufficient social pressure issuance can be increased to benefit certain stakeholders. This would be catastrophic for Ethereum.
20 Likes
Is there a link to the proposal? I can’t seem to find an open PR on the eip repository.
Edit-1: I miss understood that this will be alongside eip1559. Will take a second look to better understand this. The below comment has yet to be adjusted. This is not a counter proposal
Thank you!
It seems that this EIP doesn’t quite target similar goals. If I’m not mistaken, one of the goals behind 1559 is to make gas prices more predictable (not necessarily lower). This seems to not affect that whatsoever.
I’d be keen to see a ~counter proposal~ that also attempts to make gas more predictable as that also benefits the wider community.
Looking forward to seeing this transpire.
2 Likes
I strongly disagree with this proposal.
I seriously doubt the PoW chain will last for 2 years (eth devs seem to want to accelerate the merge now). I think fears of ETH getting 51% attacked are way overblown (I seriously doubt a decrease of hashrate of like 20% would negatively affect the network), and a PoS merge is happening sooner than you might think.
7 Likes
NACK - this breaks a very well established social contract of minimally viable issuance. Not only would this tax users and holders to the sole benefit of miners, on a user first network, it would set a precedent where by miners may make similar tantrum-like threats for any change in the future that MAY impact their short term bottom lines over clear improvements to the protocol. Ethereum is a user focused network, not a miner focused network.
Horse trading with miners is not necessary. The dynamic of mining market competitiveness, the function of the difficulty algorithm, and the strong contingency of social coordination to mitigate attacks by miners are strong enough deterrents that I believe any such “shows of force” are merely theatrical, not a threat to be taken seriously given the motivations under game theory.
The “threat” this proposal attempts to address is a reduced security posture due to an exodus of miners who can no longer remain profitable. But in fact it isn’t. It’s an obfuscated ultimatum. The cost to attack Ethereum via rented hash attacks if miners leave due to negative short term profit impact is an irrelevant threat if 51% of the network can already mobilize via a show of force to a) enforce their will on governance via fear, or b) actually attack if so inclined. Scapegoating rented hash attacks is silly if your next move against 1559 is to prove coordination of 51% of the hashrate is possible.
The only counter proposal acceptable to such a threat, if made real by colluding hashrate, is an abandonment of now verifiably flawed PoW
13 Likes
Can you explain how 2 years was decided as a mitigation time?
The EIP’s motivation section talks about the risk being “A sudden drop in PoW mining rewards could result in a sudden precipitous decrease in mining profitability that may drive miners to auction off their hashrate to the highest bidder while they figure out what to do with their now “worthless” hardware. If enough hashrate is auctioned off in this way at the same time, an attacker will be able to rent a large amount of hashing power for a short period of time at relatively low cost and potentially attack the network.”
All of the above focuses on short-term security issues from the addition of EIP-1559’s burn model that would be implemented instantly one block to the next - the ‘cliff’. Yet EIP-3368 is trying to help smooth out the reduction in block rewards over a 2 year period. However, there is no substantiated reason why this would be a 2 year problem for Ethereum.
On top of this, the block reward sounds to be scheduled to “decrease slightly every block” indicating a linear decay. Meaning for the first 12 months issuance will be higher then a flat 2 ETH per block reward. Then the second 12 months issuance will be lower then a flat 2 ETH per block. The second 12 month period then being the time where the long term supply is ‘made whole’ within that 2 year window. The problem is the likely hood of the merge of PoW to PoS not occurring within 2 years is basically nil. This means this will almost certainty result in a net increase to issuance - leading the chain away from Ethereum’s goal of “minimum viable issuance”.
Simply put, the 2 year decay period is far out of balance with what is ultimately a rather short lived issue.
4 Likes
I believe the eip needs an update. Per a Twitter thread, the eip should also read:
Requires: eip-1559
This would reduce some confusion.
6 Likes
Minimum Viable Issuance makes sense in a Proof of Stake world where you have built in penaltiy/slashing routines to keep incentives aligned., but in Proof of Work you are purely counting on the ‘minimal issuance’ maintains enough of the potential hashrate for the network. While this has worked for the past 5+ years, this recent bullrun has seen a 2.5x increase in force projection on Ethereum and continues to rise due to Fees/MEV + existing BR. The average ‘reward’ payout over the last 4+ months has exceed 3.6 ETh. The Feeburn is expected to burn 10-40% of the additional 1.6 ETH if not more and on days of heavy activity where block rewards are upwards to 6-10, we could see more burned then the 2ETH inflation.
A planned offset with decay puts for a flat baserate plus burn with EIP-1559 included. Your still providing value to the hodl position and ensuring Eth maintains the 2nd strongest network next to Bitcoin.
This basic ‘we are making miners rich’ narrative is misguided. There are more new miners every day because the ecosystem allows for more horizontal growth in securing ETH. This information war campaign on the mining ecosystem is misplaced by the very people that live under the veil of its security.
18 Likes
Maybe PoS comes sooner, but nobody here nor the devs can say that with any definative statement. A checked in notional specification while good to see is effectively the break glass option and fix later. While it may sound the rally cry for 10 minutes when people actually sit back and realize what that could mean they will not op for a half backed solution. Liquidity Providers (which I am one of on multiple pairs) and other DeFi functions are not going to just rely on a pissed off fast merge approach to a completely theoretical implementation. Testnet, unit testing and other CI needs to shake out functionality before that really gets merged.
The EIP also needs to be updated to “OPTIONAL” if the Eth Price/MEV keeps rates where the risk to security is still low post EIP-1559. This is a counter risk if there is a large drop in overall issuance due to high basefee burn and low ETH prices. Not against compromise on the rate/decay.
19 Likes
This is not about mental gynanstics with your ‘well established social contract’. The risk is actually pretty linear. The 51% attacks are not Theoretical, they have happened on many PoW chains. The nefarious actors don’t care about your decorum or ideologies around MVI. The arrogance lead ETC to getting attacked 4 separate times before they had to implement a anti-51% methodology.
What I want to ensure is there is a real plan, implemented and ready to trigger that addresses a mass exodus. Ethereum should have that projection period. The algorithm has 0 protection against a 51% attack. Other networks with Ethash have been successfully attacked. The know how, the scripts and intention is there. The fact we have to debate is shocking. The force projection of Ethereum is nearly 2.5x ATH. This means the total network PoW GPU/ASIC hashpower is profound. Much of that hashrate is very new participants. The attack vector is there. Nicehash, MiningRigRentals, Cudaminer and about 10 other separate broker services all broker hash to the highest bidder.
I have 0 confidence ACD and this body of participants are agile enough to address the problem before it would happen, if there was a existential threat. It’s like arguing for the safety check on a plane so you can have a cheaper fair.
19 Likes
Yes, It is being proposed as a included with EIP-1559. 1559 would go in as planned, 3368 would go in as a safety net as maximum security posture prior to PoS merge. Climate has changed, 410+ TH plus nearly 120+ TH on other networks is a mountain of potential hashpower. The entire mining climate has expanded due to the bullrun as it has allowed more to participate.
Much of the viewpoints people have on mining is a ‘vertical’ revenue increase model, but in reality, horizontal (*participants) grow faster than any one single entity. It’s not one miner getting wealthy, its the 500+ new participants a day that start mining. (individual contributors) improving decentralization of hashpower against a federation of pools (40-60 world wide)
18 Likes
We can effectively model such economic security levels to provide a reasonable assuredness that attacks (especially rented hash) are too costly or not costly enough to carry out. What I have not seen is any modeling of the potential loss of hashrate 1559 will result in, if any, and what that means at various price levels for the cost to procure enough hashrate to attack post 1559.
But I’m not convinced regardless. The network was secure in the DeFi 2020 Summer with lower total fees in USD terms, much lower ETH price and 200k GH/s less hashrate. Miners come, miners go as the market ebbs and flows.
If rentable hashrate is a viable threat under 1559, other proposals make more sense than paying miners more in what amounts to extortion. Fast merge to ETH2, hard fork to an ASIC friendly algo of a lower cap coin, etc.
6 Likes
And folks bet the farm on this assumption. This is not 2017 #s; why would people even want to assume a vulnerable algorithm
9 Likes
Minority hashrate chains, and ultra low caps. This isn’t new information. It’s literally the game theory behind PoW in action. Majority hashrate chains do not suffer as high a likelihood of this risk. The only way to completely mitigate it is to fully abandon PoW.
7 Likes
You have not provided any evidence or any models supporting your claim that a 51% attack will happen if EIP1559 passes. Meanwhile, there has been tons of evidence from the other side. The facts are that hashrate was 20% lower than it is now just one month ago, and no attack happened. If an attack was feasible then it would have happened last year when defi was popping off yet the hashrate was less than half. This is just blatant fearmongering from miners.
7 Likes
My statement was looking past the recent glass break option you noted. Given recent events I 110% understand why that would be the primary thought, but truly I was more-so referencing the ‘normal’ timeline where the merge can is released without the quick / fast merge approach. It seems your first paragraph was geared towards that. Which is appreciated the response, but I’m not sure struck the heart of my concern / question.
I’ll try to narrow down — by the time London is released w/ 1559 included the Beaconchain will have been running for 6 to 7 months. Assume the price / MEV / fee rates show a security risk and 3368 is launched, this means a 2 year decay window puts us to mid-2023. What is the basis for thinking this will be an issue going into mid 2023?
2023 is also notable because even with the old (and now defunct) Phase 0 / 1 / 2 model the docking was still expected in 2022. Given changes to that model to eliminate Phases all together and the general sentiment to merge before sharding (even talked about outside the scope of 1559) it seems going into 2023 is essentially not going to happen and even a late 2022 is unlikely. Even the docking page of the Ethereum website is guessing 2021/2022 - https://ethereum.org/en/eth2/docking/
I understand track records / our lack of crystal balls can not lead us to a definitive statement, but surely a mid-2023 phase out is very conservative for a merge date. Mid 2022 puts us 1 year out from London and would be a more reasonable estimate for a merge date (even including the testing you noted).
So what I’m getting at is 2 years just seems far too long for something like this and doesn’t seem based on any meaningful metrics. All of which looks beyond the point I hit on initially - this is a security issue that is relatively short term (think months, not years). As you noted, a cliff. So the merits of smoothing out make sense but the decay period seems very far off. While my opinion is meaningless on the grand scale of things, I would wager a majority of the community would see 2 years as a non-starter. I don’t see how this gets traction at all unless that window is at least 1 year or less… likely even 6 months or less. The 2 year length makes the compromise of a 1 base reward essentially meaningless to the broad community since it will potentially get cut out completely or more likely just shortly lived. I wouldn’t even begin to support this until the decay was at the 1 year or less length.
1 Like