11th Old Friends Reunion Talk: EigenDA: Converting Cloud To Crypto

Sreeram Kannan, CEO & Founder of EigenLayer

IOSG Ventures: Now let’s welcome our very distinguished guest, Sreeram Kannan, Founder and CEO of EigenLayer. And he’s going to present us: < EigenDA: Converting Cloud To Crypto >, thank you!

Sreeram Kannan:

Thank you so much to the IOSG Ventures team for arranging this Summit, I’m really excited to be here. Given that we had a great introduction to EigenLayer already. I’m going to talk about EigenDA and the theme that we explored. When and where we build EigenDA, how do we bring Cloud to crypto?

Okay, so you see the logo here within the clouds, you know, some of you may recognize it. How are we doing this? When you think about the era of Rollup, you may think of two different objectives. One is how do we outsource the Ethereum L1 traffic to our Layer 2? Whereas that’s not our thesis, our thesis is how we bring cloud-scale computing to crypto.

So firstly why do cloud apps need crypto? So if you think about two, different kinds of apps, there are two fundamentally different axes. So value per bit is how valuable the application is per bit of transaction and throughput which is what the rate you are communicating, and crypto apps today are operating at high value per bit, low throughput per bit, but cloud applications really rely on exactly the other dynamic where there is a lot of throughput but with low value per bit.

Each tweet that somebody writes, how much value can that have? But a lot of them together add a lot of value to social networks. For example, why do cloud apps need crypto? Crypto brings Native incentives, User governance, and Permissionless innovation. The last one, is one of our favorite topics: Permissionless Innovation; How do we ensure that anybody who has great ideas can build on all of the existing applications? If you are building on top of an API or Twitter or Facebook or something, then you’re always worried about when is this API going to shut off, and when am I going to be internalized into the core protocol.

But in crypto you have rigid immutable verified API that can be built on top of it, but cloud apps also need very high throughput. You know we cannot make do with the amount of throughput we have today, we also need very low cost per bit so these are the two dimensions that EigenDA really focuses on.

Before we go to that let’s first examine Why Rollup?

So when you think about cloud and consumer apps, there are several dimensions that don’t fit the other deployment models that we have in crypto today.

The first one is the User Experience, you want the instant user experience, you want super-fast confirmation and what can be done better than a single central sequencer?

This sequencer is kept in place by a variety of checks and balances where the state transition functions

correctness is checked by a validity proof or an optimistic proof, as well as censorship resistance, is brought in by ensuring the transactions in return to Ethereum L1 or directly to a data availability layer have to be included, but the instant confirmation changes the UX completely.

The second one is when you want to bring Cloud-native applications. We cannot ask everybody to write their programs in the EVM, we need new virtual machines, new programming languages, game engines, AI inference engines, all of these natively into our blockchain infra. So that’s something that Rollup really helps us do.

Another thing that I think very few people have understood is that when you have a single central sequence you can do something that you cannot do with a decentralized blockchain which is subjective admission control. What is a subjective admission control? You know we want to have no fee, free usage for authenticated users, this is not something that you can do on a blockchain, because you know a blockchain has to ensure that there is no spam, and the only way to ensure that there is no spam is to impose a price, and a price is a very non-discriminative mechanism, it cannot distinguish between MEV Bot, that’s just spamming the chain and a real user who actually wants to use your application.

So that’s something that we could do with Rollup. Why can we do that with Rollup? Because we have a single central sequence or the sequencer can exert subjective admission control. For example, if you already have a Facebook or Twitter ID, you should be able to log in and use the application without paying any fees, even without having a wallet, this massively simplifies the user onboarding experience, so Rollup has superpowers that few people have understood and finally, we want to get all of that without losing the benefits that we are used to.

Composability, with other applications that are building on top of the blockchain space. We want to get existing liquidity, we also want to build on other developers’ work. One of the things we are trying to do with EigenLayer is to ensure that people specialize in building really resilient systems without having to do redundant work. And if you look at the large arc of civilization, it bends towards each of us doing more and more specialized things and consuming more and more generic things, and that’s really where we see Rollup and EigenLayer play a role here.

Okay, but Rollups are great, but they have a lot of problems. What are the problems that we see with Rollup

The first problem is the throughput.

So if you go talk to any crypto dev you’d probably be worrying about block space, you know, how do I have access to block space if my demand for block space is going to go up, is somebody else going to come and like flood the block space? like you know Yuga Lab, does their next bored ape drop and suddenly you’re drowned in traffic and you can’t get in.

But this is not the case with the cloud, the cloud space auto-scales, the cloud has more space if you have more demand so that’s how crypto should be, but it is not.

Okay, Cost economics. If you’re a cloud developer, you are used to very guaranteed performance and cost basis

crypto devs face high and erratic costs, even if the cost is low you don’t know when the blockchain blockspace is going to get full and you will face congestion costs.

We need security, there are ways to solve the first two problems, get high throughput and low cost. While giving up on security that’s just no good and finally we need new features. We need to be able to build new VMs, we need many more integrations.

How do we solve all of these problems with EigenDA, I’m going to wave quickly and run through these slides. The high-level idea is that EigenDA is an order of magnitude scale over anything out there. Ethereum 4844, which is coming up with the Dencun upgrade is tens of kilobytes per second, EigenDA is going to launch at 10 megabytes per second, that’s I think more throughput than what you know we know how to use today. But the next generation of developers will know how to use this scale of throughput while EigenDA has ordered the magnitude improvement over the crypto scale today. We don’t think this is sufficient, our goal is to convert the cloud to crypto which means we need much more scale, so we are working on it,

One of the things is in crypto, nominally decentralization works against scaling if you want more nodes to participate, you need to reduce the node requirements and so on. So decentralization works against scaling. EigenDA scales horizontally which means EigenDA decentralization is scaling. The more nodes you have, the more throughput you can pump through the network, no single node ever needs to download all the data, that’s the architecture of EigenDA.

Okay, so what about the economics you know, the random price fluctuations and so on. So if you look at a Rollup and compare it to something like Layer 1, today you see that Rollups are not really that competitive with Layer 1s in a few dimensions.

The first one is the cost of DA, data availability which is where you write your data to a common data repository is pretty high; the second one is data availability costs are uncertain, even if it is low today, in a 4844 launches and you’re like yeah I know it stands next lower, so let’s just go. You know build all our applications there and then somebody comes up with some new inscription thing that’s going to drown your entire bandwidth.

My gut feeling is that 4844, even though it’s bringing more bandwidth is not going to reduce gas costs that much, because of this uncertainty, it’s a big problem whereas if you’re an L1, you don’t have this problem, you know exactly how much the cost basis is, because you know nobody else is interfering with you.

Finally, Rollups take exchange rate risk which means if Rollup had a native token and they’re paying fees, they don’t know how much the fees are fluctuating, because you know the Rollup token to ETH may fluctuate, and so there is an interchange risk.

Okay, but for a Layer 1, that’s not the case, because you can just give a portion of your own inflation. How do we do, what do we do with EigenDA, Rollup can outrun Layer 1s.

Why is it? The first one is the data availability costs low, we have plenty of data availability built on a hyper-scale system, so costs are low. And the second one is we have the ability in EigenDA to do long-term reservations, just like you go to AWS and reserve an instance that only you use and nobody else touches, you’re reserving a lane of data availability just for yourself, that’s possible with EigenDA.

Next, even when you’re making this reservation, you’d nominally pay ETH but you could also pay in your own native token which means you fix a certain inflation of your native token that actually is being used to run data availability, finally, roll up tokens can be used in EigenDA for dual staking.

So what it means is the data availability system is protected not only by ETH Stakers, it’s also protected by your own token committee. All this work is outsourced to be managed by the EigenLayer and EigenDA systems, so that’s what EigenDA does for Rollup. EigenDA brings Ethereum-centric security, this is something that the architecture of EigenDA is the Ethstakers participate in EigenLayer, they can restake into EigenDA, we also bring the decentralization from the Ethereum node operator set to EigenDA, and finally, we say we allowed the rollup tokens to be used natively for dual staking.

So all of these things ensure that we have high data availability security when you’re building on top of EigenDA, and finally, there is beyond the data bandwidth.

There are other limitations that you see with Ethereum today, for example, the finalization time is too slow, it takes 12 minutes to finalize a block, there are new blockchains coming up and saying like I’m going to

give you confirmations in a second or lesser not competitive.

So what can we do, we can build new services on EigenLayer to actually solve these problems, you can have a superfast finalization layer like the one NEAR is building, and you can also use for example, decentralized sequencing layers like Espresso to do this thing. Instead of relying on centralized sequences, you could use decentralized sequences instead of bridging which is rickety and low-trust models.

Can we have bridges that are really rigid when somebody wants to move data from one L2 to another L2, you have enough stake in backing the bridge and you go out and immediately accept the receipt on the other side

and move value around, so you can have near-instant confirmation using restake bridges finally, you have like much more powerful MEV management tools that you can build with Eigenlayer that you can hook into

if you are a Rollup on EigenDA.So EigenDA brings a lot of auxiliary services to EigenDA Rollup.

Now I mentioned some of the limitations of Ethereum, but also there is additivity which is there are new oracles, there are new keepers that do even driven actions to factor authorization networks, fully homomorphic encryption, ZK proof verification, AI Coprocessor, and many many new categories that will make it really easy for Rollup that are working with the EigenLayer ecosystem to integrate overtime, so with that, I would like to conclude my talk.

Thank you so much for IOSG Ventures for putting this together.

Youtube Video Link: https://youtu.be/YDP6mvcxwdg