podcasts
Collection
Collection for podcasts • 511 items • Updated • 2
text stringlengths 15 3.02k |
|---|
Speaker A: Welcome to bankless, where we explore the frontier of Internet money and Internet finance. This is how to get started, how to get better, and how to front run the opportunity. This is Ryan Sean Adams and I'm here with David Hoffman. And we're here to help you become more bankless. You know how we say blockch... |
Speaker B: I think if there's any sector of conversation that you and I really just love is the intersection of cryptography and economics, of numbers and economic manifestations on Facebook. I love protocols. That's our love language, this episode. Is that so? I mean, we've talked about EIP 4844, we've talked about pr... |
Speaker A: Yeah. So going through the implications is perhaps my favorite part. And this is one of my favorite types of episodes that we do because it's both frontier and imminent. I mean, we're just talking months away from, from this upgrade. So it's right around the corner. It's very relevant to our here and now liv... |
Speaker B: This is like the near term stuff. |
Speaker A: Yeah, this is stuff we get to look forward to. So, guys, we're going to get right to the episode with Domothy. But first we disclose. Nothing big to disclose today. Both David and I hold ether. You already know that. |
Speaker B: Shocker. |
Speaker A: We are long term investors. We're not journalists. We don't do paid content. There's always a link to all bankless disclosures in the show notes. All right, we're going to get right to the episode with Domothy. But before we do, we want to thank our friend and sponsor, Bankless nation. We are excited to intr... |
Speaker C: Yeah, I'm happy to be here. Thanks for having me on. |
Speaker A: All right, so we're going to get into what blobs are blob space, how it's different from block space. And I think we want to do that in kind of three different sections here. The first is history. We're going to talk about how we got here and why rollups are really going first. And then we'll get into the te... |
Speaker C: Yeah, it's pretty much all correct. |
Speaker B: Before we get into the. The history, Dom, can I just ask a very simple question? Is to understand blob space. Is blob space just block space for layer twos? |
Speaker C: Pretty much, yes. All the data that layer twos need to commit on chain are going to go into these blobs, which is the new resource, as Ryan said, into layer one. And the layer one doesn't know what's inside these blobs. It's just there to prove that layer two committed that so that no one can cheat. |
Speaker A: So that's all it is. Blobspace is just block space for layer two s. And right now, layer two s are actually using Ethereum block space instead of blobspace. And what we're doing with this new upgrade, EIP 4844, is we're partitioning off this new resource called Blobspace and we're making that change cheap an... |
Speaker C: Yes. |
Speaker B: So Dom, in order to fully understand how we got here, how we got to blob space, I think it's worthwhile going back into memory lane to understand the fullness of the Ethereum roadmap, because it came to a very logical conclusion of blobs and blob space. So maybe we can go back in time, because at one point i... |
Speaker C: Yeah. So in the research space, it was always assumed that the solution to scaling a blockchain was going to be some form of sharding in one way or another. And back when we were at proof of work, it didn't really work to split the blockchain into 64 or I 1024 or whatever power of two number of mini blockcha... |
Speaker A: So Dom, when you say sharding, that's basically just splitting ethereum into different pieces, right? |
Speaker B: Or any space. |
Speaker C: This is concept that already exists in computer science or database management. Basically it's sharded as in a lot of smaller shards of the blockchain get much more scalability in parallel. So this is the, that was the goal, but it turns out it's pretty hard to do that safely with execution charting. And it ... |
Speaker A: Yeah, I was going to ask you, and this is, by the way, we're going to memory lane. So this is somewhere around 2019, I think just about, yeah, that was around the timeline where we sort of discovered that, oh, sharding. In this way, full execution layer sharding is going to be really hard and it might overly... |
Speaker C: It was a very big upgrade that was planned, if you remember, around 2019, that was the ETH 2.0 slogan, where the goal was to have one big upgrade with proof of stake, with sharding, with everything along, we were going to. |
Speaker B: Go from the stone age of Ethereum to Sci-Fi Ethereum in one single upgrade. We were going to get all the upgrades that we've gotten in the last four years at once. |
Speaker C: Yeah. So one big jump and it turns out that's hard to do. There's a lot of complexity involved and it was split up into different phases. And then we said, okay, first we're going to launch a beacon chain, then we're going to figure out how to actually merge it with the current execution layer. And then we'r... |
Speaker B: I want to keep on going and defining execution charting just because once you understand execution sharding, the current roadmap of Ethereum, the roll up centric roadmap of Ethereum is placed into even better context. I think a quote from your article, Dom that you wrote, is that sharding? Execution sharding... |
Speaker C: Yeah, exactly. So I would say that the way we're getting execution charting this way is more indirect with roll ups and data sharding. But it's kind of like a cheat code from a research perspective, because Ethereum layer one has much fewer things to do and worry about, and then the rest is offloaded to roll... |
Speaker A: There's so much, I guess, to unlock here and there's a lot of, I know we're trying to explain blob space at the 101 level, but there's so much, I guess, stacked knowledge that feels almost necessary to explain. Going to this episode. We have a ton of resources at bank list to explain these, these various top... |
Speaker B: What's true being the order of the blocks. |
Speaker A: Exactly. The order of the blocks, exactly. And the processing of those blocks. The next layer out is the data availability layer. My rough approximation of what that is is what's happened. You've got what's true consensus, what's happened. That's the data layer of a blockchain like Ethereum. And then you hav... |
Speaker C: I would say you got it pretty much correct. Hints this more on a certain point is that data availability right now is more implicit and it boils down to trustless verification. We want everyone to be able to verify the chain by themselves and not have to have a trust me bro a third party in the middle, which... |
Speaker A: So let's talk about that for a minute. So the rollups post the Ethereum data availability layer so that it's not a trust me bro type of situation. Exactly what does that mean for a user inside of a roll up? Does it mean they always have a way to withdraw their assets from the roll ups? Or like, what kind of ... |
Speaker C: Yeah, so at its core, we say that rollups inherit the security of l one, which is a very quick tagline for rollups. The way it works is you have the layer one contract that bridges asset between Ethereum and rollups. Ideally, this smart contract can and should in the future all be immutable and not upgradabl... |
Speaker A: This is why, by the way, I think Dankrad said recently that his definition of an Ethereum roll up is a roll up that posts its data on Ethereum. And if it doesn't do that, then it's something else. It's not an Ethereum roll up. Maybe we use the term validium or something else for that. Do you subscribe to tha... |
Speaker C: Dom, he actually said this about layer two, not roll ups. The roll up term is very specific. That is what he means, he says layer two are roll ups and there's no other. There is no layer two thing that is in the roll up. According to Dankrad. He means that as in, like you mentioned, validiums that have data ... |
Speaker B: Not only do you have to rely on them, but you also have to rely on the mechanism, right? So you could post your data on bitcoin if you wanted to, and bitcoin is super trustless, but then you need to trust whoever's facilitating that relationship. Typically we call that a bridge. And so I just wanted to add t... |
Speaker C: Roadmap anyways, the idea is that rollups can afford to do sacrifices that layer one can't do, because layer one enforces constraints on layer two with the bridge contract that I mentioned earlier. So it's something that we can't just achieve the scale and the order of magnitude that rollups can. Even if we ... |
Speaker B: Isn't this related to the dynamic I was talking about earlier, where execution sharding is like the state sponsored version of sharding, where every single shard is totally homogenous. They're all the same size, they all go at the same speed. And then what you're saying is like, well, and Vitalik is saying i... |
Speaker C: Yeah, it's related, but the quote about Vitalik was really about scalability, because even an EVM roll up is going to be much faster than a sharded layer one. |
Speaker A: One thing I find interesting about this roll up centric approach, which hasn't been mentioned, but I think kind of implied with what you were saying, David, is the original version of Ethereum was very top down, central planning. It's like the Ethereum protocol is expanding main chain in all of these differe... |
Speaker B: Is not developing rollups. |
Speaker A: Yeah, and there's something interesting with the analog that we so often use of, like, nations and how they're built. And the first model is, how much should a government actually do right? It's kind of like an open question in different societies, I feel like, have different answers to this question, but th... |
Speaker C: Yeah, it's a very astute observation, but in terms of being intentional on that, I would say it was kind of forced upon Ethereum researchers because many points we've already touched on is execution, charting is complex. And the way we're doing with the sharding plus roll ups, it's kind of a cheat code, but ... |
Speaker A: Okay, so that's a little bit of the history and how we got here. And I think, I hope we explained, gave you a tour de force on how we got here to a roll up centric roadmap. Now take us to the current state, if you will, Dom. So right now we do have many roll ups. And right now, these roll ups, these ethereum... |
Speaker C: What's wrong with it is that a few cents is still too much for scaling blockchains worldwide. Like Vitalik said years and years ago, the Ethernet Internet of money should not cost $0.05 or $0.50 or the exact quote. And it's something that people make fun of him because of layer one gas fees. So the current s... |
Speaker A: So it's too expensive, it's limited, and we've got this weird kind of resource issue wherever. Yeah, coupling resource coupling issue. That's a great word for it, David. And so what happens is 100,000 users on a roll up are competing with 200 NFT bidders on main chain, and they're all competing economically ... |
Speaker C: Yes. So the resource coupling problem is kind of a weird thing economically, but there's a technical reason, historical reason, why ethereum went to couple everything into a single unit of gas. But it's weird when you think about it, because these 200 NFT bidders are using execution mainly to send money and ... |
Speaker B: So not only are these resources coupled, but what you're saying, Dom, is that these resources are inappropriately coupled, as in they don't have to be coupled. One is using data, the other is primarily using execution. And these are two different resources that just because everything is contained inside of ... |
Speaker C: So you want the technical answer because we already end waved it. Blob is just a piece of data that roll ups can put a onto main chain. |
Speaker B: Yeah, I think it's time to get into a technical. Okay, so I'll ask again. What's it technically? What's a blob? |
Speaker C: Technically, a blob is 4096 field elements that are just under 32 bytes each. So that's the, that's the. The playground that roll ups have to put data and they have to abide by this format. |
Speaker A: What is the difference between a block and a blob? |
Speaker C: The blob contains basically everything that trade everything at execution. And I don't know how to explain this. It's like it contains the transactions, and then you execute the transaction, and you see that I sent you one eth, and that changes the state of ethereum layer one. It's like I have one less eth a... |
Speaker B: Wait, so we have blocks and then we have blobs, and then we have data of transactions. Transaction data goes into block space. Blobs go into blobs space. And everything is contained in a block, correct? |
Speaker C: Yes. |
Speaker B: The new thing is this new blob space. So we have block space. And I think people might get. And I'm getting confused, but I'm trying to parse this apart, that we have blocks and we have block space and we have blob space. But blob space and block space are spiritually equivalent, and both of them are going i... |
Speaker C: Yeah, so we're getting into metaphor. Land on the consensus layer. What we say is that these blobs are some sort of sidecar alongside the block, so the block can stay small. And then there's the big blobs next to it that are not exactly contained in the block, but the block contains a reference to the data, ... |
Speaker A: I think one of the first learnings here, the first takeaways in this section, because it's going to take us a few minutes to really understand what the heck a blob is and what the heck blob, blob space is. But you know that thing that we say on bank lists so often is blockchain. Sell blocks. Well, I think, D... |
Speaker B: Dom, you said that a block references a blob, but that a blob isn't in a block. Is that correct? And can you elaborate? |
Speaker C: Yeah, that's the metaphor. The blob is the sidecar that's alongside the block. And then basically one key insight is that, like I said earlier, if I send one ETH to you on layer one, every single node has to compute the transaction to verify my signature. Verify, access the state of your account and subtract... |
Speaker B: Right? Okay, so the way that I understand, I give a metaphor to blockspace is that block space is like a container of data. And if I send Ryan some ether, that's a very small bit of data that I'll throw into blockspace, start to fill it up. If I send Ryan an NFT, that's an even bigger bit of data. If I mint ... |
Speaker C: It's similar, but it's a decoupled market too. So there's this other different bucket that only gets filled with data. Like the layer one doesn't care. It's very agnostic about how who posts what into these blobs. All it cares about. It's getting these blobs out there for whoever needs them to commit to them... |
Speaker B: Okay, so post EIP 4844, when we introduce blob space, aggregate Ethereum blocks, the size, the data size of aggregate Ethereum blocks will be blobspace plus block space, correct? |
Speaker C: Yes. |
Speaker A: Okay, let's look at this through the different lenses of actors and stakeholders in the ethereum ecosystem. So I want to go through validators in a second, but let's first start with rollups. Why is blobspace a better product than blockspace for rollups? Is it just because it's cheaper? Or what other propert... |
Speaker C: From the point of view of a roll up sequencer, it really boils down to being much cheaper and much more plentiful. Layer one is going to do more technical stuff to scale this blob space, but from the point of view of a roll up, it doesn't really matter. It's just data for them. That's all they need and that'... |
Speaker A: And how do they sort of activate it? Is there anything that they need to do on their side? Is it basically like, so all of these proofs that they were posting and buying block space in order to prove they just switch over? So it's just like, rather than using, I don't know, electricity for your furnace, you'... |
Speaker C: Yeah, exactly. So after 4844, they'll have to update to support these blobs and stop posting everything on layer one in the expensive call data section of blocks. |
Speaker A: Is there anything that they'll lose by posting doing this via blobspace versus block space? Or is it just basically an equivalent substitutionary good? It's just as good for roll ups. |
Speaker C: It's just as good and even better for them because it's cheaper and more plentiful. |
Speaker B: Can I put a transaction in blobspace? Can I send Ryan some ether in a blob? Or what prevents me as a layer one user from consuming blobspace? |
Speaker C: Nothing prevents you. It's a permissionless system. It's tailored for rollups. But of course you can put any data inside layer one blobs as you please. But it's up to you if you want to pay for that. |
Speaker A: And I'd be more expensive. Maybe a bad idea. |
Speaker C: Yeah, I'm predicting that there will be some NFT project where the jpegs are inside blobs at first because blobs are going to be pretty cheap. |
Speaker B: Wait, okay, so there's a phenomenon in the NFT world about like nfts that are on chain like cryptopunks and mfers and a few other NFT projects. What's the famous one, the generative one that auto glyphs are like the biggest on chain art? Does blob space just simply allow for more? Perhaps as one use case of ... |
Speaker C: Yeah, I'm pretty think there will be a lot of Degen stuff happening with blob space because it's going to be very cheap at first. |
Speaker A: Oh no, we don't want that though. We want the blob space for our roll up transactions. We don't want to create another Degen market. Not the point of resources decoupling. |
Speaker C: Yeah, it's decoupled because if they want to put data on chain, we can't stop them. Right. They could just pretend to be a roll up and we can't. Yeah, we can't be top down about who gets to use what from a credible neutrality at layer one. And I don't think there's a way to enforce that to that. To enforce t... |
Speaker A: But probably the properties of blob space are such that they are most conducive to, well, it's like, isn't there, so one example, Dom, isn't there like an expiry on the blobspace? And this brings us to we'll talk about the expiry. But this brings us to the second lens I wanted to talk about, which is from a ... |
Speaker C: Yeah. So with 4844, it's going to go up a little bit. You have to download and serve these blobs all the data that's inside the blob, but you don't have to store it forever. So that's what you mentioned. Blobs will expire after, right now the specs say about 18 days, I believe, and after that it becomes like... |
Speaker A: So this might be a reason why it's not as conducive as block spaces to an on chain NFT, although there may be ways around this. But why is this 18 days not a problem? So why isn't that a problem for layer two is like if the data goes away after 18 days, aren't I like, yeah. |
Speaker B: Wait, pruning data away from blockchains is blasphemy. How dare you? |
Speaker C: Dos. |
Speaker A: Yes. How could you? |
Speaker C: So this is kind of the weird thing with the name data availability, because it's not data storage and it's not continued data availability forever. So it's kind of the opposite. There's a big problem with storing everything by every node forever that doesn't really scale. If you were to pay once to get your ... |
Speaker B: Okay, so data availability, in contrast to data storage availability, as in Ethereum, is making this commitment to the world around it, that it is making data, it is enforcing. That data is sufficiently available for the surrounding universe to be able to take and do what they need with that data in order to... |
Speaker C: Yes. Yes. So the goal is to really have lightweight nodes be able to verify everything, including the availability of what's been published in a scalable way. And also, another thing to point out is that this blob data is not part of the state where you need to read and write to it many, many times a second.... |
Speaker B: Right. Notably, terabytes of data are under $100 right now. So maybe if I'm. Maybe we can theorize that in the future, there will be some software application that you run on your desktop computer, and you link it to your main addresses, and then it follows your addresses around some pre selected layer twos,... |
Speaker C: Yes. And also rollups can have their own designs to incentivize storage of their own data if it's something that's very important to that particular roll up in its community. So the design space is interesting. |
Speaker B: So in addition to pushing execution out to the free market and making layer two teams optimized for execution, we're also just pushing data storage out to the free market as well. |
Speaker C: Yes. So one metaphor that I've used, Vitalik uses that Ethereum is supposed to be more of like a billboard, where you can have information in real time about what's on the billboard, but once it's erased, all you can do is verify that this information wasn't on the billboard, available to everyone to see bac... |
Speaker A: Yeah, that's kind of the image I get from Ethereum as well. It's kind of like a present moment of time type of consensus mechanism. It's not trying to store the full consensus truth of the universe, the full computer. It's more like kind of like a. |
Speaker B: Hurricane or just kind of like tip of the chain. |
Speaker A: Yeah, it twirls around. And so that is this distinction between data availability and data storage. So Ethereum is trying to be a data availability computer, but not a data storage computer. And so this 18 days is also significant because it does have to be like, there does have to be some reasonable time pe... |
Speaker C: Yes. It basically just has to be long enough for anyone who needs that data to download it. That's why I said earlier, it's kind of a cheat code for execution charting because then you have these roll ups posting blobs, and if it's a roll up you don't care about, then you just don't download the blob. You're... |
Speaker A: Just to tie off this question, you said the requirements for a validator do increase a little bit. |
Speaker C: Like what are we talking about with proto Dank sharding? I did the math in my blog post. With the current size of a blob and the target number of blobs every block, we're looking about 50gb extra storage by every node. But this is for 4844 after full denying it's going to go down somehow, blobs are going to ... |
Speaker A: The goal of ethereum in general is to be able to run a validator on consumer grade hardware. Hardware, right. And we meet that goal right now. Like you can run a validator on a raspberry PI. And what you're saying is this blob space changes the hardware requirements by about 50 gigs of hard drive space. So d... |
Speaker C: Not truly your wife. |
Speaker A: Yeah. How will they be able to do it? |
Speaker C: The typical user? Hopefully all the technical stuff is abstracted away from them. It's more the heavy users who really insist on having trustless verification of everything to secure their funds themselves with their own keys. And this is what blobspace allows them to do. They just keep that data, only the d... |
Speaker A: View, yeah, I guess this will just accelerate the migration from Mainnet to roll ups as well. I'm sure we could probably predict that this economic change will result in that. |
End of preview. Expand in Data Studio
README.md exists but content is empty.
- Downloads last month
- 4