Ethereum Cat Herders PEEPanEIP, 13 Apr 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: PEEPanEIP #31: EIP-3198: BASE FEE opcode with Ratan Rai Sur

Description

Topics:
1:21 - Introduction
2:08 - EIP-3198 Overview
4:00 - Motivation & use cases
6:15 - Futures with Ethereum Gas, is it similar to buying gasoline at low and selling it at high?
8:30 - Do these futures already exist?
9:29 - How is BASE FEE opcode more precise than the gas token?
11:50 - Specification
12:17 - Would this proposal make no change in the BASE FEE but make it easily accessible?
13:11 - How does the present chain work in absence of this proposal?
13:50 - While submitting a transaction, the wallet (Metamask) suggests a gas price, how to derive the gas fee, is it available on the block, or somewhere? Why did 1559 introduce the BASE FEE?
15:50 - What is the primary change in estimation of the gas fee?
17:00 - Does the BASE FEE exist on the current mainnet?
18:00 - What do you mean by a mint gas token?
20:15 - What is the challenge period?
21:50 - Difference between BASE FEE & gas observability? During client implementation, did you find any edge cases that suggest making gas prices accessible to the EVM is really a security concern?
24:15 - Why burn the BASE FEE with 1559 transactions?

Resources:
EIP - https://eips.ethereum.org/EIPS/eip-3198
Discussion - https://ethereum-magicians.org/t/eip-3198-basefeeopcode/5162

Follow at GitHub or Twitter:
Ratan Rai Sur @RatanRaiSur1 | Alita Moore @AlitaMoore85 | Brent Allsop @Brent_Allsop | Pooja Ranjan @poojaranjan19

Peep an EIP #16: EIP-1559 with Tim Beiko - https://youtu.be/ZbSKo_x9RvU

Ethereum Berlin upgrade community call - https://medium.com/ethereum-cat-herders/ethereum-berlin-upgrade-countdown-community-call-ffa51239c148

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

Follow PEEPanEIP earlier episodes at https://www.youtube.com/playlist?list=PL4cwHXAawZxqu0PKKyMzG_3BJV_xZTi1F

Contact Ethereum Cat Herders
---------------------------------------------------
Discord: https://discord.gg/tzYmDmF
Twitter: https://twitter.com/EthCatHerders
Medium: https://medium.com/ethereum-cat-herders
GitHub: https://github.com/ethereum-cat-herders/PM
Email: support@ethereumcatherders.com
Website: https://www.ethereumcatherders.com/

A

Welcome back to another episode of penny, an eip dedicated series where we discuss an ethereum improvement proposal and understand how it is going to make ethereum even better. My name is pooja ranjan and I am your host for today's episode on base p opcode eip3198 basebr code is a standard track core proposal created in january 2021.

A

It is currently in review status. According to the eip. This proposal adds an arc code that gives the evm access to block space fee. This proposal is authored by vitalik buterin abdel hamid.

A

This is a companion proposal to eip 1559, free market changes, and that's why I also expected in london soon to talk about the proposal. We have gratin, sir, also known as rai from the consensus quorum. Bazoo team. Welcome to the show rathan. You have joined us for the first time as a guest. So for our audience, uh if you could, please introduce yourself and tell us a little bit more, how you got engaged with ethereum.

B

Sure uh yeah, so, like puja said, uh my name is rai. I work on the hyperledger basu ethereum client, currently employed by consensus, and I got interested in ethereum.

B

Basically, when I was traveling in 2018, I was living in brazil learning portuguese hanging out on the beach, and I had nothing better to do really, and I discovered this thing called ethereum and I started reading about it and it just made a lot of sense why a system like this should exist, and so, when I got back to new york, I just uh started teaching myself blockchain concepts implementing like some toy protocols, myself and implementing papers and then uh yeah applied, applied to consensus.

B

um Okay, so erp 3198 the base fee. um This is a simple proposal um and it requires uh 1559. So there's some background for 1559 that I think is valuable to go into.

B

So let me swap over there real quick. um The main thing. um Let's see.

B

Here we go yeah, there's a basic, so basically, the main thing that is applicable to the base vr code is the base feed network wide parameter in 1559., so as big blocks or small blocks will be added to the 1559 enabled chain the base fee will dynamically adjust in order to increase or decrease quantity demanded uh for uh for the block for transaction for inclusion in the block and therefore represents a kind of measure of congestion of the block.

B

um What it actually does to the protocol is that it measures how much uh or it dictates how much of uh how much each should be burned per gas, and so, as you can imagine, big blocks will make that base be higher and small blocks will make that base lower and we try to stay around a target of 12 and a half million gas um so because the base fee represents the kind of congestion the network making it observable in the evm also gives that that knowledge of congestion to uh contracts within contract execution.

B

um The implementation is so simple, because the base b is already known uh in the block header, and so there's not much to do. Except for just push this value of the base fee onto the stack so that subsequent op codes can manipulate it, but it's worth going through kind of the motivation.

B

What use cases does this enable?

B

So this first point uh contracts that need to set bounties for anyone to poke them with the transaction could set the bounty to be base b plus x or base fee times one plus x. This makes the mechanism more reliable because it will always be enough, regardless of market conditions, so this kind of makes sense, because if you had a bounty that was lower than the base fee well, then why would you send a transaction in order to retrieve it?

B

um The dependency on the base fee make sure that this is always is always a viable, or this always makes sense, as a transaction gas futures can be implemented based on it. This would be more precise than gas tokens, so gas tokens is a contract that kind of represents um how much you know, state you've stored in this in this contract, and you get these gas tokens that represent that and then because ethereum has refunds.

B

You can then, later on, um you know, send these gas tokens to the contract in order to free up a lot of storage and potentially, uh essentially, you have an arbitrage opportunity between times when uh gas prices are low and gas prices are high.

B

um The reason you can implement gas futures with the base fee is imagine you mint these gas tokens when the base fee is low. So now the contract can keep track of at what base fee. You minted these tokens um and much like futures. If you have, you know a future that entitles you the right to buy gas um of a higher value, then those will be more valuable. uh So if you mint gas tokens when the network is not that congested, these are not going to be as high value gas futures.

B

But if you meant these gas tokens when the network's very expensive, then um these are, these are more valuable. Gas features uh that you can use later on in also high uh high base fee environments, improved security for safe channels, plasma opti rolls and other fraud proof driven solutions having the base fee as an input allows you to lengthen the challenge period automatically. If you see that the base fee is high, so definitely.

C

B

C

Sorry, I just wanted to clarify one thing: yeah, um the future is with the gas. Are you basically, like I mean? Is it the same thing as in like going out and buying a couple barrels of gasoline like when the price is low and then using it like whenever it's high again just to clarify.

B

Yeah, I think I I think it's it's essentially that um so maybe an example. As far as I understand right, let's say you um mint a gas future using this gas features contract um and the base fee at the time is 200 right, so the contract kind of stores that you minted these futures with this, I don't know what they call it in the future context, but, like let's say, a strike price right with the strike price of 200 or something um in a future contract execution.

B

You are not going to be able to kind of redeem those gas tokens uh for the uh in in a high base fee environment, so, for example, you're like gas tokens that were meant, so your gas futures are minted at 200 right, they're not going to be valuable. If the environment the contract is executing in is a high base fee environment because the contract won't, let you take them out um so yeah.

B

Basically, it's it kind of entitles you to be able to get your transactions in during the same kind of environment that you meant to them.

C

And would that cause problems? um I mean because I presume that the base fee changes to limit um network congestion and to like reduce the um opportunity or try to discourage like um over edge for gas right so like if you were to kind of like circumvent that, would that cause problems.

B

I can't I can't say for certain: I think that um the the kind of uh side effects with 1559 uh for me- I don't understand well enough to be able to, like you know, judge those interactions. um Actually, oh yeah. I should mention that abdel is the main author on this eep um and he was the one who implemented 1559 on besu and I'm just filling in for him because he's going to have a kid yeah.

C

um That's totally fine, um I'm! I guess what I'm I'm curious about, because, like do these already exist like these futures.

B

So not futures, but these gas tokens exist and uh they're kind of they don't care. What uh how congested the environment is, that they get minted or used, um and so what happens is that you mince them uh because they're completely fungible right, you mint them. When the um uh you know when gas is cheap and then you spend them when gas is expensive uh and I believe the efficiency gains come gains come from the fact that now they're not completely fungible, so uh I think you would have to mint the gas futures.

B

um You know at a certain point um in like a in the same kind of congestion environment that you'll end up being end up using it. I think.

A

Okay, thank you. I have one question related to it here, like you mentioned that this proposal that has advantage of the base fee op code over the gas token for uh looking for the gas futures, and uh I was also trying to understand how is base v op code, more precise than gas token.

B

Yeah, I think it. I think it comes mainly from the fact that gas tokens right now are fungible and uh therefore there is this kind of opportunity for uh using them when the price is low on the price, low and spending them when the price is high, um whereas gas futures you can be.

B

If you have the gas futures that are that is used, that is for only a kind of low congestion environment, um then you won't be able to use them in a high congestion environment, and so therefore there's not as much of this uh kind of you know, filling up state just to just to retrieve it later. On yeah.

A

B

um Okay, so right, okay, so the third point um improve the security for state channels, plasma, opticals and other fraud. Proof driven solutions having the base fee as an input allows you to lengthen the challenge period automatically. If you see that the base fee is high, so this is pretty self-explanatory if you know what it's talking about. So what it's talking about is that there are these off-chain kind of scalability solutions and what they require is.

B

Ultimately, you commit something to uh you know a more secure, less scalable base layer like ethereum net, um but there's an opportunity there for one party to defraud the other, and therefore there needs to be a challenge period for fraud proofs to be given uh if contracts have visibility into the base fee and therefore the network congestion, they can lengthen that challenge period if the base fee is high, so that the kind so that the kind of security guarantees are security.

B

Properties of um this challenge period uh respond dynamically to that uh and are more in line with. Perhaps uh you know time uh as opposed to block time time in like the blockchain space time in the real world.

B

um So those are the motivations and the specification is quite easy. uh You take uh when you call base fee, it's a very low cost transaction and you just push one item onto the stack, which is the block base fee and then whatever op codes come after. That can manipulate that.

C

That's pretty much it so I just want to um kind of, I guess clarify one thing: yeah, basically, the gas fee, this eip wouldn't change anything about the gas fee or the way that ethereum works. um If, if I'm understanding correctly, it wouldn't change anything um all it would do is make it more easily accessible. Basically,.

B

Yeah, so it yeah it doesn't change how, for example, yeah users will not. It does not change how users um specify their fees on any more than 1559. Does you know? That's all the same. um What it allows is these different patterns of of of contracts, um because, but we might as well make the base fee observable, so it does kind of it. It changes the uh it changes the space of possible contracts that you can create by making this inseparable.

C

It seems a little bit like um like how do contracts do it now or how like dart, or can they just not? Is it like determined after the block is mined.

B

Yeah, I don't, um as far as I can tell, there's no way to tell in your like during your transaction execution, how congested a theorem is, or anything like that you were just executing and you could be executing in a block that is nearly empty. You could be executing in a block. That's really full! You have no idea. So that's why this introduces kind of a bigger design, space.

C

I mean so, I definitely understand like the application of it, um and that makes total sense, but, like I know, when I'm submitting a transaction say on metamask or whatever it gives me an estimate for how much the gas will be um or how much the cost of my transaction will be, and so I'm curious like how it's done right now.

C

Where um can you kind of like derive the gas fee? Is it available on the block? Is it like, because I understand that the opcode itself is like a much more readily available and executable like way of handling the the gas fee, but this one really kind of changes. Things.

B

Yeah sure so uh this specific question is about, or or like the the motivating kind of example is you're. Creating a transaction and your wallet is providing you a suggested gas price right, you're wondering how that suggested. Gas price gets there right.

B

So the way that gets there is actually one of the motivations why 1559 was created, which is that uh what these wallets are doing is looking at recent um history uh of you know the recent blocks and seeing how much people are paying um in order to get included, and then there are these complicated estimation, algorithms that they're using on top of that in order to estimate it.

B

But these estimation algorithms are not so efficient, um they can often cause you know people to overpay, and that's that's one of the reasons why 1559 uh introduced this base fee, um as kind of like a in protocol way to suggest to users. Okay, this is about what you ought to be paying in order to get into the block now or within, like one or two blocks um yeah. It was completely implementation. Yeah.

C

So so then, I guess my question would be: is the primary change that you're able to see now um the price of your transaction before you make the transaction and currently like before this is implemented?

C

You can only tell the price of your transaction after the transaction. Is that a good is it? Do you think? That's accurate.

B

So the well, the kind of I think this is kind of straightening more into 1559 territory, where kind of the base fee, the basically opcode, assumes the existence of 1559..

B

So if we're talking about what uh the base, what uh you know 3198 does um is that from the from the user's perspective of submitting a transaction, if you, if you didn't implement 3198, it wouldn't be any different right. So you as a user, if we didn't implement the base via code, would still have the benefit that 1559 gives of this. uh You know easily of this base fee that you can observe and submit the transaction because you're outside of the protocol.

B

But if you wanted to observe it within the protocol, you would need the base fee. Op code.

C

So then, um like I guess currently on um main net, does the base fee exist.

B

Nope not so space yeah, yeah, okay,.

C

A

There's not even there's.

B

Not even a concept of base fee right now on mainnet, so that's why it requires 1559. um 1559 will introduce the base fee and then this will. Even this will make sense like this, wouldn't do anything today.

C

That makes total sense, okay, yeah. That makes sense um all right. No, I think I'm good. Thank you. Cool.

D

One one way that I've explained 3198 is that it brings uh an oracle native to the protocol, um whereas most oracle systems are outside and operated by people externals of the protocol. This just makes it a native element.

B

Sorry yeah, it's a very, very specific kind of oracle. Yeah.

E

Yeah, I have kind of a bit of an unrelated question, but what do you mean by mint uh gas token? Is that are you providing ether to buy that or or yeah? What does that mean.

B

Yeah, so when you mint a gas token, so what that means is so how gas tokens work is: there's a gas token contract out there um and when you mint something what you're doing is um you are just storing a bunch of state in that contract and you, when you store a bunch of state in that contract, the contract will mark that you own a gas token and a gas token corresponds to like a certain. You know, maybe hundreds or you know thousands, maybe of like um slots of storage.

B

uh So now you get this gas token, then uh you mean to that. When uh you know prices were low and then, uh when you spend it, you basically send it to the contract. The contract frees those uh freeze, those um those slots. You get a refund, so it subsidizes, whatever transaction you're doing in the moment and that gas token gets destroyed then so it's essentially a way to you're.

B

Using this uh the the contract, that is the gas token contract, is kind of like a bank or a battery um to you, know, store state, get tokens or destroy tokens and uh release the state.

E

But you use ether to mint. Those with right is what happens to the ether right you use ether to create a token yes,.

B

You use use the ether to essentially store all of the because you're paying you're paying a good amount of gas in order to store, uh essentially these arbitrary values in the storage. um So that's that's the main form which you pay. You pay ether now in order to store the um in order to store random values into into the storage um and then just like a small amount in order to do the bookkeeping uh in the gas token contract.

B

um But it's not like a the minting, isn't like a trade. Necessarily, it's not like you're, sending ether to the contract that then uh it gives you a token back.

A

So I have a question uh related to the point number three I mentioned in the motivation section that improve the security of state channel, plasma, octanes and other fraud driven system.

A

So if I understand that's a kind of help to mitigate the attempt of dos attack by strengthening the challenge period, so what is this delayed challenge period or challenge period?.

B

Right, so I was talking earlier about how, uh because these kind of octane, these specific blockchain scalability um solutions, require a commitment to the uh you know to the more less scalable more secure base layer. um This is a moment where someone can attempt to you know, send this uh or attempt to you know, submit a fraudulent state of affairs of what that kind of, for example, that state channel.

B

What that um uh you know what was the state of that, uh and so in order to mitigate that, uh you need to have a period like in which it is not yet valid, and it gives anyone an opportunity to submit a proof that shows that this actually could not be the state that an honest um actor would um would attempt to commit, and uh because you need, you need enough time to do that. It's worth it to change the challenge period. According to network congestion,.

A

That's interesting um moving on to the security consideration. There are some people who are concerned with making gas prices accessible to the evm, believing it may impose some risk.

C

A

Can cause cost future gas cost change in the future. So I'm curious to understand the difference between the base fee and the gas observability.

B

Yeah, I actually can't remember how what the gas price current opcode um returns in the case of a 1559 transaction in the case of a normal transaction, the gas price. So the gas price operate was the existing kind of observability of how much a person ended up paying for or how much the transaction ended up paying when it was included and yeah.

A

So what I'm trying to understand here is like, uh while uh implementing um this in your client visu, did you find any kind of edge cases that suggest like this actually happens, and that is actually a security concern if we allow opcodes to be accessible by evm.

B

um No for security concerns. What making things observable to the evm, what the drawbacks of making things observable to the evm are is that they become dependent on for the execution of certain transactions, and then it can potentially be more of a headache later on. Would we want to deprecate things or change, how things work or um something like that?

B

So it's not a security consideration um and there's no obviously no plans to dedicate base. We haven't yet implemented it, but it's conceivable in that in the future. um This op code, maybe wouldn't make sense in a certain context, and therefore it's behavior changes and we've seen over and over many instances where um you know and opcode is uh we kind of have to like step around and understand how a change that we propose would change the behavior of an op code. What what you know, what um what contracts that would break etc.

B

So I wouldn't consider it a security uh problem as much as just a.

B

As much as just another thing to depend on a potential like a protocol design issue somewhere down the line.

A

Right: uh okay, yeah, if I may go a little after the proposal like 15, 59 sure. First of all, uh congratulations uh for merging 3198 in beso! um I.

C

A

uh The test net for london is up now so yeah.

B

Abdel abdul really deserves all the praise for the engineering work on 1559. Of course, tim did so much of the or was the main driver of the project stuff, but yeah abdel implemented 5059 and base b yeah.

A

Yeah so yeah I mean like for people who don't know we were hoping to have adele here, but uh going on you know some leaves, so we could not have him, but thank you for joining uh ryan norris.

A

I have a small question related to that, like this, this misconception of uh burning e with uh 1559, so I, as I understand this, is basically the base fee that will be one burnt right. Why burning base b.

B

Well, there's a couple of things: one: it makes sure that it kind of enshrines ether as the way to pay ethereum in order to execute your transactions, so the ethereum protocol, if it requires a base fee, means that you can't get around requiring ether to uh be included into the blockchain and because one way to think about it is don't think about it as burning but think of it as paying the ethereum protocol. And then the ethereum protocol never pays anyone else back.

B

um That's kind of how uh how you can conceptualize burning and it's important that it's burned and not given to the miners, because if it's given to the miners, then they can coordinate in order to artificially reduce the base fee and therefore capture more of the fees for.

A

Themselves, that is very interesting. Thank you so much for that. Well, uh this was really great. uh Thank you, rai for joining us today and talking about uh the proposal.

A

um I think this is one of the fastest moving proposal towards inclusion and standardization, as you mentioned, that a lot of work being done by timbuko and abdel hamid. So many congratulations to author and, of course, the client team implementers for this great work and coordination um to me. It seems like it has a lot of useful features for also upcoming projects um like optimistic rollups and others. So I hope to see this proposal on the main net with eip1559 and summer this year, best wishes for you, devnet.

D

A

On this note, for our youtube viewers, thanks for watching, we did an episode on eip1559 with chinbyko. So if you have missed watching the overview of eip1559, that is the fee market change. You may follow the recording at evening youtube at ethereum categories.

A

Ethereum blockchain is expecting multiple upgrades in 2021. The first one is the expected on april 15th and ethereum cat holders are organizing a community call inviting everyone to join us to follow the upgrade with uh ethereum developers. The eip authors, whose proposals are going in editors and other community members.

A

The details are provided in the description section as always, we'd love to hear feedback and improvement suggestions do let us know which eip would you like to see in future episodes?

A

If you like this episode, hit like and share, we have many more interesting proposals coming up for london as well, so stay tuned, keep watching and keep sharing your love with ethereum cat headers. This is all of us signing.

A