EIP-1559: Fee market change for ETH 1.0 chain

EIPs
,
103

I personally oppose the change as written for personnal financial reasons.

There current system incentive miners to mine transactions with the highest gas price first. This lead to games where the daily or weekly price (based on blocktime of last reward+delay) is awared to the first better.
So it s a race for being mined first within the same block since everyone pay enough to get their transaction mined inside the same block.

I recently invested $5000 for running my own full Ethereum archival node in order to beat everyone in terms of latency and based on my understanding, transaction ordering within the same block would become random which would break those games/smart contracts and thus my investment.

104

Your proposal would increase the cost of running a full archival node. The 500Gib per month growth of today isn t free.

105

This is still the case with 1559. The only thing that changes is that if you don’t care about within-block ordering, then you will likely just pay some static fee. If you do care about within-block ordering you can pay a higher gas price and get that still.

Off Topic

Just so you know, an archival node won’t help you win gas price auction/front running/backrunning games.

106

And for dapps/games where participants needs to race for being mined first, there s no possibility to upgrade as the change would break the current predictable transaction ordering inside a block, correct?

107

There are no guarantees for transaction ordering. Right now it is entirely dependent on the miner. In the future maybe that will change, but there are no solid plans on how to do that at the moment. 1559 will not change the transaction ordering logic in a significant way that I know of.

108

In the case of the game I m playing, things are a bit more difficult. One need to also look at past confirmed transactions of the day for calling the contract with the right parameters.

Currently, if someone broadcasts a transaction fee higher than yours, you cancel your transaction by rebroadcasting it with a higher fee than your competitor using nonce.
As of course, other will do the same so the aim is to maintain being the highest bidder as long as possible until blocktime is reached.

Each call to web3/Etherscan involves a 200ms latency in addition to the latency of the network. A time during which your competitor is still the first. Of course, a local http node makes this better but ipc involves an even lower latency.
So it s about detecting and replacing as fast as possible.

109

Correct, but under the current auction system, it s more computationnaly simple to mine the highest gas price first and then taking the next highest price and repeat until blockgas is reached.
This act as an incentive which is why there s a de facto transaction ordering inside a block.

As far I understand once the current transaction type are no longer accepted (leaving only the new one), there would be not predictable within block ordering. I also assume if the 2 transaction types are mixed that there would be no possibility to influence the transaction ordering between the types If I m not the miner.

110

This is incorrect in the latest version of the specification (not merged yet). There is talk of deprecating legacy transactions eventually though.

This isn’t true. Block ordering with 1559 (all versions discussed) is still gas price order. If you want to be at the top of the block, pay a higher gas price. Nothing is changing on this front.

This is entirely up to miners. I don’t know what the clients are planning on implementing if we go with 2 transaction pools (a version of the specification that I don’t think anyone actively wants), but it isn’t a consensus concern and miners can change the behavior if they want (as they can now).

Transaction ordering is entirely at the whim of miners, who sometimes defer the decision client developers. Prior to 1559, transaction inclusion was entirely at the whim of miners, but 1559 makes it so inclusion is partly miners, and partly consensus. However, as long as something is valid to include, miners can still choose to include it and how to order it just as they can now.

111

How about a quadratic increase of the 1559 benefits, as per 1559Q? EIP-1559Q: Quadratic burn of gas fees as fees increase

1 Like
112

It s completely up to miners. But mining pools have to follow what pay the most and cost the least in order to compete.

My fear is if the current underlying incentive changes, miners would no longer have to perform as they does today.

113

The miners would always earn more from higher prices under EIP-1559Q just decreasingly so. Miner would still be incentivized to handle higher paying txs. The Ethereum network token would increasingly benefit as they do so.

1 Like
114

As I mentioned on today’s AllCoreDevs call, it would be great to get another round of feedback from client developers about EIP-1559. Last time it was discussed, the main issue that was brought up were concerns about the network’s ability to handle full blocks under 1559 (i.e. blocks twice as large as our current block size). We’ve been working on trying to answer that question as part of the implementation wor.

It would be good to know if there are any other major red flags that clients see with the proposal so we can proactively address them as well.

115

Here is a community outreach report on EIP-1559 that the Ethereum Cat Herders conducted: https://medium.com/ethereum-cat-herders/eip-1559-community-outreach-report-aa18be0666b5

2 Likes
116

Biased report. Lets say a miner creates an EIP that exponentially increases gas costs for Yield Farmers. Then they create a fake report that says the community is over 60% in favor if you remove the yield farmers. Brilliant!

Here is what it will sound like.
If we exclude yield farmers, 60% of projects were generally positive

Enough is enough, no more coming here and insulting this community and developers with made up “community” numbers. Show your data of 60% approval versus 40% miners ONLY not in favor.

The miners are tired of Defi claiming only they are the community that matters.

2 Likes
117

I think that for many years the job that the miners done was good enough, but now as they want to switch to PoS, they want to shit on them. I think Ethereum would be never grown so big without the mass amount of miners. I just really hope that, if they will screw over the miners, they will most of them leave, allowing the 51% attack to be easily conductible.

118

I support placing tx fees into a miner pool that is deducted against at a fraction of the accumulated tx fees as suggested by Vitalik recently. As some devs have stated, I also agree its important to state the case outside of “Oh, no, miners will lose revenue!” So this is my attempt to flesh it all out.

This tweak proposed by Vitalik regarding 1559 keeps miners incentivized to mine the chain during crypto valleys. Miners who have many liabilities have mined in the past because they expected better returns in the form of tx fees in the future. EIP 1559 as it stands, burns fees and erases that long-term incentive to build or maintain infrastructure earlier than anticipated. This is fine on its own as miners expected this for some time. However, it doesn’t replace it with another plausible security mechanism (i.e. PoS). It merely cripples the existing one. If there is another crypto winter before ETH is full 2.0, there is no reason for miners to keep the lights on this time, and hash rate will plummet or be concentrated into the hands of the chip makers.

Furthermore, this “flattened” tx fee payout adds an additional incentive that many says GPU coins lack when compared to ASIC chains - an incentive to remain on ETH long term despite having other options to mine. If there is a spike in tx fees, we can expect ETH to payout more over a longer period of time, and this will increase net security over a longer time. I think it is important to have GPU’s be the preferred choice of mining hardware because it allows for the broadest group of participants (among other reasons). These hobbyists, students, businessmen, etc become the companies, evangelists, and developers that are building on and awareness of Ethereum. If no action is to be taken on ASIC’s, EIP 1559 with a flattened fee structure will help ensure that ETH will be the most attractive (read: secure) blockchain with consumer grade hardware, and help keep mining decentralized by providing a larger and more stable stream of income despite the lower earnings created by specialized hardware players Ethereum’s yellow paper pledged to defend against. I think most miners (myself included) would champion this EIP without question if ProgPoW were implemented, but there are just too many hands in the cookie jar and its squeezing honest mining companies, even those with the most competitive power rates on the market.

If EIP 1559 is implemented as is and burns tx fees, miners will feel the squeeze from the developers and from ASIC manufacturers, there will be no reason to hold out for a better future as we know ETH 2.0 is coming, and all that will remain competitive mining ETH are the ASIC chips that have priced GPU miners out of business. I fear this is bad news for the future of ETH as it consolidates voting power into fewer hands who may be hostile towards ETH’s future plans. Every GPU miner I have met online or not is not hostile towards this larger shift.

As I have said many times around the interwebz; let’s not dismantle the old house before the new one is built.

And finally, while I appreciate the effort being put forth, I also think that Tim Beiko’s summary on EIP 1559 disregards miners as an ETH stakeholder and that kinda sucks. I get that we are squeezed in this proposal so its politically expedient to write us off, but I think conversations around EIP 1559 should capture our discontent accurately, so that our concerns can be fleshed out as we are best positioned to explain what negative, unintended consequences the mining industry (read: security) would experience if economic policies are shifted. At the end of the day, it comes across as silly to me to light piles of money on fire as a result of more economic activity so that we can hope that the money in our pockets is worth more afterwards.

I hope that these concerns stoke further conversation on this topic.

1 Like
119

Miner revenue being tightly coupled with transaction space demand results in miner payments being more volatile as demand is volatile over time. By removing that factor, miners will have a more stable fee revenue that isn’t impacted positively or negatively by demand.

When demand is high, we don’t need more hashing power, and when demand is low we don’t want less hashing power. By removing that coupling, we can ensure that we have a more stable amount of hashing happening, and thus a more stable amount of security.

If we believe that we need more security, then we can increase the block reward to incentivize more hashing (and thus more security). If there is a belief that we don’t have enough security at the moment, then someone can create an EIP that proposes increasing the mining reward.

2 Likes
120

Hi Micah! Thanks for responding. Vitalik proposed a smoothing out of fees being paid out to miners, addressing your concern as well. It ensures ETH has the highest has in its class of mining hardware (GPU’s) which ensures no 51% attack, even if ETH’s inflation rate is lower than a future competing chain. Furthermore, if miners are paid 1/1000(0?)th of the accumulated tx fees/block, it incentivizes security over the long run despite other negative factors (some addressable) that are harmful for long term network security. I think it is bad economic/monetary policy to contract and expand the money supply based on something as volatile and multi-factored as revenues of miners measured in fiat. For example, rollups will likely remove many on chain tx fees even before EIP 1559 is implemented, so using today’s fees as a basis for reducing mining rewards even further seems disingenuous to me.

121

This only solves the selfish mining attack vector, it doesn’t solve the problem of more stable miner payments unless you smooth it out over like 2-4 years. In such a case, just adjusting the issuance rate directly seems to be far simpler, and even more predictable, than trying to distribute it to miners. Not to mention, the code for adjusting the block reward is way less complicated than trying to distribute fees to miners over time.

122

Can someone explain to me why this only applies to ETH 1.0 chain? Could it be the people supporting it are large ETH holders that plan on staking in ETH 2.0 and would never support a reduction in payments for POS? Please clear this up for the POW miners. Why are we supposed to take a cut when it will not apply to POS?