Home
Contribute Contact Us Browse all meetings
Home Contribute Contact Us Browse all meetings
Ethereum Cat Herders / PEEPanEIP / 12 Dec 2022
Ethereum Cat Herders / PEEPanEIP / 12 Dec 2022
Previous Meeting Next Meeting
⏯
Add meeting Rate page Subscribe
youtube image
►
From YouTube: PEEPanEIP #92: EIP-3651: Warm COINBASE with William Morriss

Description

Topics covered
----------------------------
0:23 - Welcome to the episode
1:04 - Guest’s introduction
1:28 - Presentation
1:30 - About Gas
2:00 - EIP-2929:State Access Lists
2:47 - EIP-2930 : Warm Access Lists
3:39 - EIP-1559 Allows Conditional Coinbase Payment?
5:26 - EIP-3651 makes the Coinbase Warm
7:15 - What was the main reason for proposing this proposal
9:03 - Warm and The Cold Cost Explained
10:00 - Why accessing Coinbase is overpriced
10:40 - Client implementation & testing
12:35 - General feedback on EIP process
13:40 - Next Step towards the making Core Proposal even Stronger

Resources:
-----------------
Proposal - https://eips.ethereum.org/EIPS/eip-3651

Slides - https://docs.google.com/presentation/d/1lOpxyrdLYVKeGkudW7lUn9Vu2OVhmO4ZAJVpx1_vBRg/edit?usp=sharing

Discussion - https://ethereum-magicians.org/t/eip-3651-warm-coinbase/6640

Shanghai Playlist - https://www.youtube.com/playlist?list=PL4cwHXAawZxpok0smGmq-dFGVHQzW84a2

EIP-2929 & EIP2930 - https://www.youtube.com/watch?v=qQpvkxKso2E&list=PL4cwHXAawZxqu0PKKyMzG_3BJV_xZTi1F&index=72&t=1s

EIP-1559 - https://www.youtube.com/watch?v=ZbSKo_x9RvU&list=PL4cwHXAawZxqu0PKKyMzG_3BJV_xZTi1F&index=76&t=1s and https://www.youtube.com/watch?v=AC1FS3LmoT4&list=PL4cwHXAawZxqu0PKKyMzG_3BJV_xZTi1F&index=55&t=15s

PEEPanEIP - https://www.youtube.com/playlist?list=PL4cwHXAawZxqu0PKKyMzG_3BJV_xZTi1F

Questions for Guests - https://docs.google.com/forms/d/e/1FAIpQLSdsmujnH_ncIlSEJ23s9jzXAKwfpTGNeJqjREWPloBbkJVQTA/viewform

Check out upcoming EIPs in Peep an EIP series at https://github.com/ethereum-cat-herders/PM/projects/2

Follow at Twitter
-----------------------------------------------
William Morrisss @willmorriss4 | Pooja Ranjan @poojaranjan19

Check out upcoming EIPs in Peep an EIP series at https://github.com/ethereum-cat-herders/PM/projects/2

Contact Ethereum Cat Herders
---------------------------------------------------
Discord: https://discord.io/ethereumcatherders
Twitter: https://twitter.com/EthCatHerders
Medium: https://medium.com/ethereum-cat-herders
Website: https://www.ethereumcatherders.com/

A

Increased accessibility at lower cost as one of the primary objectives of improving ethereum blockchain transactions. If you want to learn about a proposal to form coinbase continue watching.

A

Episode 92 a series dedicated to ethereum Improvement proposals where we have an overview of an upcoming proposal and protocols for ethereum Locked In I am Pooja ranjan from the ethereum casualties, and this is a special episode on Shanghai upgrade. We will learn about eip3651 warm coin days created in July 2021 EAP 3651 is a standard track EIP, which is currently in reduced status and is also selected for Shanghai, upgrade looking forward to learning more on the proposal in this episode with our special guest and the author of The Proposal William Morris.

A

Thank you William for joining us today.

B

Yes, so I'm an engineer: I've been in the space since 2017. I'm, introducing warm coinbase to provide an improvement for people that use conditional payments so.

A

Thank you for sharing this with us without further Ado, let's peep in.

B

It's a short EIP, but we'll provide a lot of background. So uh guess is how we meter the computer that is ethereum and mainnet has for context. 15 million gas per blocks come every 12 seconds, and so a simple transfer would use 21 000 guests, whereas in a more complex transaction might use the entire block.

B

So when we talk about gas amounts, we're talking in the context of these numbers and so eip's 2929 added State access lists to the gas computation of a transaction, so the more State a transaction accesses, a distinct state that is the more a transaction will cost. Previously, all accesses would cost the same, whether you'd accessed it or not, and so effect all addresses were called, and so this EIP also defined a few accesses.

B

That would be warm from the start that is they're accessed already at the start of the transaction and they're the from account, which is the origin. The two accounts specified in the envelope and the set of all pre-compiles, which is well defined because pre-compileds can expand and we don't have to add them specifically or specify them in the future. So the next EIP 2930 allowed transactions to specify access lists. This was included in Berlin, it introduced transaction type. One and this feature is still available with eip1559 transactions, transaction type 2..

B

In both of them you can specify and access that you're, anticipating and assuming that it happens. You'll save gas, of course, if it doesn't happen, you'll waste gas. So you should only use this for accesses that you believe, with 90 percent confidence or more that's uh you're planning to access this uh storage, and so uh there's also an a Json RPC method. Eth create access list that facilitates the creation of these access lists, which would be wonderful if more front ends would adopt, as that would save their users gas. Next.

B

The last piece of background is that eip1559 allows conditional, coinbase payments. So previously we all paid transaction fees via gas price times gas and so there'd be an easy way to calculate your maximum fee. And but the new method introduced by this in London was to separate the fees into priority fees and Max fees, and so you can set the max fee to the base fee and the priority fee to zero and thereby pay the miners, nothing when you reverse, and so inside of your transaction.

B

You can pay the miner conditionally by using the coinbase op code to get their address and sending them ether, and so this allows you to pay transaction fees only when the transaction succeeds, and this method is especially popular in systems like flashbots, which I've Linked In This slide. You can access this slide in the description. It's this is the uh just a an account that receives a lot of such direct payments, and so you can examine the transactions that are engaging in this Behavior.

B

Another nice property is that if you submit a transaction that begins to revert, uh miners will quickly forget about it, because they can't keep all the noise. It would be bad for them to be constantly continuously simulating transactions that don't pay them. So usually they don't keep such transactions for long and they get censored because they don't pay the minor when they revert, which is what you want for your transactions. That revert, you don't want to pay a transaction fee. Instead, you want it to never happen um so now without final delay.

B

Eip3651 makes the coinbase warm so conditional payments previously mentioned for one five: five: nine. They need to access the coinbase accounts in order to pay it directly. This access is cold currently and I'd like it to be warm, so the the workaround that's currently being used is to sign many transactions per uh coinbase.

B

So if you expect that the coinbase might be the flashbots builder, then you can sign a transaction that adds the coinbase Builder to the access list, for example, and if you're 90 sure they'll be the builder, then uh then they have an advantage over everyone else, and so that could be bad because it results in centralization Block producers. Another thing that could be bad. If we don't do this would be, it could be that people would pay in things besides ether in order to reduce their accesses.

B

So it's just a great simplification to add the coinbase address to the initial access list. It circumvents a lot of the complexity here and the redundancy of signing so many additional transactions, and now we have a period for questions this.

A

Is a great presentation.

A

This proposal itself is like shot like 3651 dot. Mv5 I have seen. This is one of the shortest proposal. Yes, yeah I'm sure EAP editors must have loved moving it across one status to another status. Yes,.

B

Yes, so uh simple, because it refers to specifications in other eips.

A

That is right, I totally get that and yeah. uh That is my first question as well. It looks like the proposal was introduced right after deploying the Berlin and London a Creator where 2929, 28, 9, 30 and 1559 was included. You did mention that uh the there is a workaround. However, when 29 29 was introduced, coinbase wasn't the part of uh the starting access address. According to the author, the reason could be somewhere between oversight and simplification.

A

What would you say as the main reason for proposing this proposal like a significant advantage or the hassle of a work around to warm the coinbase.

B

Well, it certainly improves the simplification of the that I don't have to sign as many transactions I don't have to sign one transaction per Miner, for example. um That's the people that might benefit from this, uh though, are the people that would use conditional payments, and this would just use less gas for them. There's bigger fish to fry uh when it comes to that, though, ether transfers still use 9000 gas, and so it can be cheaper to send weth, for example, depending on your previous accesses. That's.

A

Right, thank you for sharing this um for people who may not have a fully known proposals discussed earlier EIP 2929 EIP, 29 30.. We have recorded an episode with vitalik buterin and Martin check out episode, 20 on people. The link could be found in the description and similarly, we do have episode for elp1559. We will share the link in the description for you. My next question is around explanation.

A

Maybe for our novice listeners to follow up the conversation completely. Could you maybe explain the warm and the cold cost associated with the transaction.

B

Yes, so the first time you access an account or a storage slot, it will cost more than every other time after that, within the same transaction. So from this slide, 2929 uh the prior cost for an account access is 2600 guests and the prior cost for storage is 2100, yes and uh subsequent accesses when they're warm are only 100 and so for account access. This would save 2500 guests for a coinbase tip. According to the current specification.

A

That is quite a lot and help me understand this correctly. Is it what we mean by accessing coinbase being overpriced? Yes,.

B

It's that compared to the amount of computation required by the network, this operation costs too much, and so it's using more of the block than it needs to, and so this will make things more fair in the gas market.

A

Right and the biggest Advantage would be like making it more fairer.

B

Yeah, the main thing that gas costs should be is fair because we can make all gases zero and just abandon it all together, but then it wouldn't be fair. uh The network needs to be used by many different people uh using it. Many different ways.

A

Awesome and my next question is related to implement ing statuses. We are aware that is proposal ify for a Shanghai by any chance. Are you aware of plant implementation, or maybe, if you are aware that it has been included in any of the Shanghai.

B

I haven't seen any progress on any of the forks and towards implementation, but implementation. The hardest part will be writing the test case to make sure that the coinbase is there. Another snag with implementation might be estimate gas which currently doesn't specify the coinbase address so geeth, for example, lets you specify a constant coinbase address.

B

So if you send eth estimate guess to that, you might be able to hard code an address that isn't uh going to be the coinbase in the uh anyway there's going to be like a few Corner cases that are implausible to encounter, except if you aren't aware of how the system works. So the uh like it's possible to trigger this intentionally more so than accidentally, people often configure the zero address by default and then otherwise one of their addresses, and so it's possible to miscalculate.

B

Therefore, but in practice, I use common miners for that and that's a better practice than using one of your own addresses if you're not actually producing blocks.

A

Right and um I just checked the execution specs of uh the folder shanghai.md, where all proposals are listed and it seems that eip3651 has been merged in every client. So yeah, that's a good news. Yeah.

B

It's they said it would be simple, and so I was never worried about that.

A

Right very well, my last question is about the EIP process, so looking into the repo EIP Depot, it looks like a EAP. 3651 is one of your latest proposals, but you have shared three other proposals in the past, which unfortunately, are in stack and status. Now so I'll be really curious to hear your general feedback towards the EAB process for core proposals.

B

um I think that uh the engineers are pretty good at evaluating engineering proposals and I'm satisfied with the amount of discussion that I'm seeing there I wish that they would engage on The, ethereum Magicians Forum more uh when they have strong veto opinions in particular that way bad ideas can get shut down earlier in the process, but otherwise I'm satisfied and impressed with the procedure. I hope that it continues to govern us first, the foreseeable future.

A

That is awesome. Thank you for nice words um in the same uh context, I have another question by any chance. Are you aware of the newly proposed core eits in an executable spec world? If you may have came across this, um what do you think? Would it be a good next step towards uh making the core proposal even stronger, like you mentioned about having more discussion in Fellowship of ethereum magician, this is more related to having it implemented directly on execution specs, rather than getting into an EAP process execution.

B

Specs is the test Suite right that.

A

Is currently under progress, It's not fully done yes,.

B

That's certainly better than uh like an implementation on only one client uh I understand that some people try to push their eips by implementing it on every client. Most people don't have the resources to do that, and uh so it's much better. If, if there's a simple way to help the core developers- and it seems that I believe you're referring to the test Suite is that is that correct.

A

No, it's actually a python voice, converge specs,.

B

Okay, yeah I understood that they have a test suite and that might be a good Common Place uh that everyone uses it sounds like more of them are adopting that I'm not familiar with the execution. Spec, though.

A

Not a problem I hope uh next year we would be learning more about it. Shanghai was supposed to be an experimental upgrade in which we would be running both EIP process and execution specs process in parallel, but there could be some delay. However, we'll be talking with all the EAP authors uh to collect their feedbacks.

B

Yes,.

A

All right well, um thank you. So much for sharing all these things, uh it's almost time to wrap up. Is there anything that you would like to add or share with ethereum client Developers.

B

uh No thank you. William.

A

uh We appreciate you taking our time to share about CFI proposal, eip3651 warm coinbase and with the release of this episode, my hope is that it helps the community, especially the tab, development Community, to share their support for this proposal, and we see it included in Shanghai a place that is expected next year. On this note, thanks everyone for watching and listening to Shanghai special episode. Should you have any question, let us know in the form available in the description or leave a comment to reach us at each other.

A

This is gone check out, description for links to useful resources and guests Twitter to follow we'll be back with another interesting proposal. Overview till then keep watching keep sharing your love with ethereum Caterers cheers.