Welcome to East to E to E to E to E. All right, our next talk is what the is Serge from David Leonardi from Nevermind. Thank you very much. Hello. Hi, Stam.
I'm David Leonardi. I'm a senior project blockchain uh senior blockchain project manager at Nethermines. Um, and my talk today is about surge, which is a based roll-up templates that one of my teams at Nethermine has been working on for the past uh half a year plus. Let's talk about roll-ups. Um, they say that roll-ups are just fancy multisigs that they don't inherit from Ethereum security and that they are slow.
So today we'll look at these claims and identify some solutions to some of these problems. Let's begin first of all by understanding what uh roll-ups are and why they matter and what the current solutions are and how they fall short. So Ethereum's base layer can only process so many transactions due to its decentralized nature. So rollups has emerged as a solution to batching transaction offchain and then posting compressed proofs back to Ethereum. This approach has been successful and increasing with throughput.
But it comes with a critical design choice and the basic concept is simple. So instead of executing every transaction on the Ethereum mainet, roll-ups execute them in a separate environment and then post these compact proofs back to Ethereum. So this increases the number of transaction that can be processed while it is inheriting security from from Ethereum. Most today use a centralized sequencer model. So you have like FIFO or whatnot.
And a sequencer essentially is a single entity which is responsible for collecting user transactions, determining the transaction order and producing batches to submit back to Ethereum. And the centralized approach creates several serious issues. So first it introduces a single point of failure. If the sequencer goes down, the entire rollup just stops processing transactions. Second of all, it creates transaction vector um censorship vectors.
The sequencer can choose which transactions to include and which to ignore or delay. Um third it concentrates power in the hands of one entity which can extract value through transaction ordering similar to me issues that we see in layer 1. Finally it relies on reputation and feature decentralization promises rather than trustless mechanisms and this is where the concept of based roll-ups comes in. So no it's not just a theory u and the term based actually refers to based on ethereum. So uh leverage uh Ethereum's existing infrastructure uh for the critical function of transporting of transaction ordering rather than introducing a separate and centralized component.
Um a based rollup takes a fundamentally different approach by eliminating the sequencer uh the centralized sequencer entirely. Instead it relies on Ethereum's own validators and block builders for transaction ordering. In this model for instance uh users will submit transactions directly to the L2 memp pool uh like here um and then searchers will come around and uh search for profitable bundles of transactions that can be sent back to the builder uh which then orders the block and passes it to the proposer. A compressed state route is then passed back to the L2 and gets picked up by L1 searchers. sorry is passed down to the L1 and gets picked up by L1 searchers which then gets included in L1 block again via the same PBS mechanism which is ultimately validated by the L1 validators.
So it's fundamentally a different approach and instead of creating a separate sequencing layer, we're using Ethereum itself to handle transaction ordering and this is closely tied to the proposer builder separation uh PBS right uh which is a concept in Ethereum's road map uh that aims to keep B build block building open and permissionless think of it as Ethereum native of sequencing. So essentially we're leveraging the existing infrastructure rather than building a parallel system. Um the benefits of the system are profound. Um so first there's no centralized entity that can censor or reorder transactions. So you can get a transaction into Ethereum and it will be processed by the rollup.
Second the security is identical to Ethereum itself. So base rollups inherit Ethereum's decentralization's property directly. Third there is no need for additional infrastructure or governance mechanisms for transaction ordering. You can just use what already exists. And uh lastly uh the architecture is simplified and more costefficient because less there's less overhead and less components based roll-ups are fundamentally aligned with Ethereum's values from day one.
Uh and they're pretty much as decentralized as Ethereum itself because they just leverage the same security and infrastructure. This approach really represents a return to the core principles. So decentralization isn't the future roadmap itself but it's baked in really from the beginning. Let's talk about roll-ups. What are the properties that an Ethereum align and neutral rollup should have?
A rollup should be permissionless, trustless, censorship resistance, open source, performance, and anybody should be able to operate it. Let's talk about surge. It's a roll-up template for fully decentralized based roll-ups that unlocks gigas performance and is directly stage two from launch. Um so surge is a fully decentralized base rollup template that delivers stage 2 from day one right uh and it operates with a tokenless design using only ETH. It achieves gigas throughput through the nethermind execution clients and has a two out of three multi-prover security model all while maintaining complete alignment with ethereum's core values and of decentralization security and censorship resistance.
So that's quite a bit right. Um, so let's try and break this down step by step and see why this matters and how we can create something that really innovates the roll-up space. So how do we get here actually? So we used the Tyolthia stack um which is currently the most mature production gradient battle tested grade uh based L2 rollup but we made a few modifications to it. Um so we replaced the finality gadget to remove all of their tiers and instead we use a two out to three proof system.
We replace the bond from using the Tao token to use only ETH and we change the split the fee splits to be like from 25 75% to be 100% of the base fee which will then be burned uh even though we reserve the right to change this on the mainet deployments. We'll see how that goes. And we replaced the execution client from Tao Gath to the Nethermind execution client which allows us to handle Gigas. Um, we also added extensive documentation to allow anybody to just come in and operate any component of the network or to just deploy their own search fundamentally and just try it out. Um, so for those of you who don't know what stage 2 means, let's take a look at that.
So L2 beats uh is an independent organization that evaluates the security of layer 2 solutions and they've defined a framework in which different stages of security uh are like explained and stage two is the most rigorous. Um, so stage two security includes that uh a rollup must have a 45day exit window for any upgrades, giving users ample time to withdraw their funds if they disagree with any like protocol changes. There's no emergency powers. So there's no ability for anyone to force changes without going through the time lock. Um, and lastly, there's limited governance.
So the multisig can only cue upgrades which then have to go through the 45 um like time 45day time lock, but they cannot execute them immediately. And this really gives like security to the users. This means that from day one, search provides one of the strongest possible security guarantees. And we're not asking users to trust us, but we're building a system that actually doesn't really require trust. And this is in stark contrast to most rollups where often from the start u like the foundations operate with extensive admin powers and emergency back doors promising to remove them eventually, right?
Uh but we just fixed that in stage two. Uh we also don't have any token. It's tokenless. So unlike most L2s, Serge doesn't introduce a new token and we just use ETH. So the ticker is ETH.
Uh this isn't just a philosophical choice. It actually eliminates uh any misaligned incentives that often come with their own token economics and speculative dynamics and just really issues from technical to economical and it really with uh the uh economical interests of Ethereum validators and users. Um so it also dramatically reduces overhead by avoiding token distribution, governance and regulatory complexity. So it just makes it super easy. Um next up is Gigasp.
Um so the Nethermind execution clients uh is able to handle gigas per second performance as as has been proven in the 1.30 release or 130 release uh on the base network. And um this is mainly done or this mainly happens because of like some technical optimizations and how the application is built. Um, oddly enough, it's net that actually provides some really great benefits. Um, yeah, it works.
Um, and lastly is our multiplier multiprover security model. So, um, we replace the finality gadgets and we now have a two out of three consensus model that requires that uh, two out of three proof systems must be in agreement. Right? So, we have uh, an SGX system, an SP1 and risk zero. And, um, this fundamentally eliminates any single points of failure.
So um in the proving layer and creates a higher security thresholds than single prover architectures and all of these systems can use different security assumptions and uh have uh let's say there's less attack surface because common vulnerabilities are minimized. Um the system also remains secure if one of the provers is compromised because it's two out of three and this actually provides much higher fault tolerance. So together all of these innovations uh create a rollup that's both high performance and highly secure. uh really in line with Ethereum's uh core values. Um what can you do with Surge today?
Um you can try it out. You can go to Serge Wtf and you can check out the website, learn about it. Um you can come and deploy your own. You can run your own infra. You can read about it on the docs.
Uh you can bridge in uh transact on it. Uh and just see how it is. Um yeah. So let's talk about some of the additional things that are coming uh with Surge, right? We call these powerups.
Um and these are kind of like upgrades in capability that will add more features. So we're going to add pre-confirmation soon which will bring uh transaction uh speed to like you know few hundred milliseconds tops. Uh we're going to optimize the proving mechanisms. We're going to like also improve the block building and we're going to work on uh seamless L2 composability to like have a more flourishing ecosystem. Um yeah.
So what's next for for Serge? Um we really want the community to come and engage. Uh this is really meant to be um let's say a shared effort uh through the entire Ethereum ecosystem. Um so please come and deploy stuff on Serge. Just try it out.
It's meant for developers. It's meant for builders. It's meant for uh tinkerers, right? It's not meant for end users. This is a roll-up template.
Strong emphasis on that. Um, and we're going to come and continue improving the finality gadgets and just make sure that that is more secure and more tolerant. Um, but we also want the community to come and take over this rollup template and just do things with it. So, we're not going to come and deploy yet another rollup. That's not the business we're in.
We're building a roll-up template which we want the Ethereum community as a space to just take over. So, one of these examples is Ethereum R1 which you may have heard of in the recent days. So the plan for Ethereum R1 is to use Serge as uh as a network or as a rollup template that is deployed deployed on the chain. Uh there's no governance again the ticker is E. There's no token.
It's fully community owned. It's run on donations which will directly go into funding the further development and infrastructure that will be needed to deploy and operate the chain. Um and uh after 2030 like everything's just going to be burned. um which is yeah um there is no governance there's no cult of personality like it's you know purely for the community as a research project from railers for builders um yeah you can join us on this journey please uh check out this link uh it'll take you to surge wtf where you can find uh all of the information about the project um you can read the docs they're quite extensive and again you can bridge into bridge.hoody.
surge surge.wf. You need some hoodie e there's a link to a faucet. You can get some there. Um yeah, thank you so much.
Thank you. There any questions? Yes. No. Yeah.
Thank you. Does it work? Yeah. Okay. How does uh decision making work at a communityowned project without a token and without governance?
um we are not in control of the deployed rollup that the community will operate right so it's up to the community to form its own consensus on how to operate things so uh that is not the business of nethermind to do an example what's happening right now within R1 is that there's a community that's forming that's making decisions on how to operate the chain but um since we do not provide any alternative token there's you know no misaligned incentive of well there must be like token where number go up it's just ETH right so most people that show up there it's it's quite interesting it's just devs that do things for the sake of doing the right thing and in that sense I'm quite happy because we inspired builders to come and build awesome tech because it's the right thing to do uh and that's exactly what's happening so uh you're welcome to join uh those conversations um and shape it myself. Hi. Um, does um search use suave uh for uh the auction mechanism? I mean it goes through PBS, right? So, it's just searchers doing whatever searchers do and then proposing that or passing that to a builder.
So what's behind that? It's really up to the builder themselves. Okay, thanks. Time for one more. Oh, maybe I could have missed it, but did you talk about pre-confirmations?
Yes. So that's still coming. So we're working on it. It's going to come in the next few months. Uh the intention there is to have 200 milliseconds performance ballpark.
That's the number of being whispered. Nice. Thanks. Thank you. Amazing.
Thank you so much. That was excellent. [Applause]
Automatic transcript — names and jargon may be misspelled.