Data Availability Summit 🗄 Intro to Modular Blockchains and Open Problems
ETHGlobal·Wed, Oct 9, 2024, 12:00 AM
Mustafa, Co-founder of Celestia Labs, will give an overview of modular blockchains and their benefits, the state of the modular blockchain ecosystem in 2023, and open problems. ETHOnline is a three week long online hackathon with over $200,000 USD in prizes as well as a series of curated summits, celebrating some of the most significant happenings & learnings from the past 12 months in web3. ----------------- Follow us: https://twitter.com/ETHGlobal​ Check us out: https://ethglobal.co​m Email us at: hello@ethglobal.com
Transcript
[Music] yeah so I'm going to give a quick introduction to mod blockchains and there's four parts of my talk I'm going to be talking about what mod blockchains are and their benefits their progress in 2023 open problems and the final destination of what we're trying to achieve with modular blockchains so between 8 and 2020 blockchains were basically fit into this sort of Monolithic era where you had Bitcoin and ethereum which had a very similar architecture where effectively you had a virtual environment sorry you had an execution environment for Bitcoin it was Bitcoin script for ethereum it was the ethereum virtual machine and some consens mechanism and those two things were kind of coupled together so all the all the nodes ran the execution environment and executed every single transaction on the chain and this wasn't scalable so in 2019 this idea of decoupling consensus and execution started taking hold and the general idea was well what if um we had some we had one chain that was only responsible for consensus and availability but the execution was offloaded of chain and that's basically how rollups work rollups they post their blocks to ethereum or some other consensus and data layer like Celestia but the actual computation and the actual execution and verification of these transactions happens offchain in the rollup chain itself and those those the execution for the transactions is proven to the consensus layer using proving systems like fra proofs or validity proofs and that basically what you end up with that is a sort of a modular blockchain Paradigm where instead of having the consensus layer and the execution layer kind of enshrined into each other like eum was when it first launched you now have this modular blockchain ecosystem where the base layer is only responsible for consensus and data and on top of that base layer you have a vde variety execution layers in the form of rops and those ex execution layers are modular because they can all kind of have different architectures and disadvantages and advantages and they can use different programming languages so for example you have non evm rollups so your roll up does even have to use the evm like for example sarware uses CYO and I'm going to decompose what the different layers are of the modular stack to discuss a little bit about the theory of modular blockchains so as I mentioned first of all you have the consensus layer and what the consensus layer does in the nutshell is it basically takes in some arbitrary messages that a developer has provided and it orders them it provides provides some ordering over them so that people can agree which messages came first because you need to you need to know which messages came first to know if the same coin is being spent twice which which which transaction should be the valid one and then on top of that you have the data availability layer which is once those messages are actually ordered we need to prove to the world they've actually been published because it's all well and good ordering those messages but if we haven't verified they've actually been published the network then no one can actually know what the state of the chain is what the transactions are and do things like prove fraud if some if there's an invalid transaction it and this is actually very similar to kind of like what the Bitcoin white paper proposed as the as the essential component of Bitcoin that makes Bitcoin possible was this idea of a timestamping server like this idea where you can just take a hash of block of items to be time stamped and Wily publish widely publishing the hash that's that's that's the main core aspect of the blockchain and so Bitcoin was kind of like on the right track there but Bitcoin itself enshrined execution environment but the core component is the time stamping server so the core inside of modular modular blockchains is what if we isolate that core component and allowed people to build their own execution environments on top of it and if the Bas layer is is only trying to be responsible for data availability and consensus then it can specialize more in trying to scale data availability in ways that are more efficient than if it tried to do everything at once and that includes using techniques like data availability sampling where you can guarantee or you can check and verif that most of the block is available has been published by only downloading a very small piece of that data and so that's a very good way to scale blockchains because now it's no longer the case that everyone has to download every single transaction and the entire data of the chain to verify it and then on top of that you have the execution layer which uses the base layer below it which provides consensus and data availability and what the Executioner does is it takes takes in those arbitrary messages that have been ordered and published and then does some kind of computation on top of it to Output a state and what that basically means is if you assume those arbitary messages are transactions then it does some computation on those transactions and then it outputs for example a table of accounts and balances and that's basically basically the job of rops in the kind of in a roll op Centric ecosystem they do this execution and they offload the data at the consensus to the Bas layer and then optionally you can also have a settlement layer which is sort of like a special case of an execution layer that basically Bridges other execution layers together so if you have roll up a and roll up B and um they can use some cement layer by by posting their proofs to it and that settlement layer can allow liquidity to flow between those two rollups and so kind of acting like as as a bridging Hub so if we put all that together what is a modular blockchain we Define a mod blockchain as basically a blockchain that fully outsources at least one of those components that we discussed so let's go through some what what some of the benefits of modularity and modular blockchains so firstly scalability because um rops now have their own dedicated computation environment which means that instead of having to verify instead of having a single chain where everyone's transactions are executed and that was kind of like the original ethereum World computer model where there was a single kind of world computer and everyone's transaction is executed on the same chain on the same same computer and that doesn't scale obviously but with rollups if you spin up a rollup rollups have their own execution environments so if a transaction happens on rollup a then rollup B does not need to execute that trans that transaction and so that's that's very similar to sharding it's effectively sharding execution and is inherently more scalable the other thing is that if you use techniques like data availability samp sampling you can create a system where the more light nodes you have then the more block space you can have that while still keeping the chain verifiable and secure because with data availability sampling light nodes can contribute to different verication of the chain the other Advantage is choice so um imagine if you had imagine if when you're building a website you only had one choice of programming language like what if every time you create a website you have to use PHP or Python and you could you can't you can't use anything else that's very similar to kind of the status sare with with the ethereum viral machine um people with modal blockchains and with the ability for people to create rollups people can now experiment with better or more customized execution environments so for example you can have a paralyzable execution environment you can have a execution environment that is more suited towards a specific purpose for example there's a project called um curio which which has modified the ethereum virtual machine to make it more suitable for like Real Time games or executing Real Time games on it by embedding the actual game itself as a up code to the evm and there's also different types of rollups that you can play play around with like Sovereign rollups or normal rollups or celestium which are sort of like um roll ups that use ethereum rollups that use year for the a or ethereum l2s I should say and so users now have a lot more choice in choosing what kind of um components they want to use in their applications and then you also have this idea of sovereign rops um so there's a very big Cosmos ecosystem but it's it's not uh viable for for every time to create a new chain you create a you validate ass set so with Sovereign wollups it kind of gives you the bre of bur worlds because you can create your own Sovereign chain without having your own validate set from scratch so quick overview of what the modular St looks like today um there's a very there there's a lot of it's very wiely wearing sack you have different projects in all different categories including the those four categories I mentioned but also things like sequencer Services Ro Frameworks even roll up as a service providers you you can just go on the website uh upload your code and immediately deploy your own rollup chain in seconds and various kind of cross chain bridging providers and so who knows what the the ecosystem will look like in a year's time um so let's go through some open open problems in the modular blockchain ecosystem the first one is kind of like we need this we need better ux for bridging because in a multi- roll system um bridging kind of becomes more of a pain for users especially if there's many different roll up and application specific rollups as well then payment systems for resources across layers so if a rollup has its own gas token and it also needs the gas token for theer the kind of needs to be better building for that to make it more seamless for people deploying rollups there's also a lot of choices so we need to make those choices is easier to understand for developers and maintaining dependencies across the stack um is also a challenge and there's this idea of maybe we should Define common interfaces for example a common da interface for all Frameworks instead of having a different slightly different one for every framework we also need to kind of like faster ZK proofing systems and better devel developed for proofing systems I don't think there's any kind of completely deployed decentralized for proofing systems right now that's completely permissionless and also privacy this time you have an opportunity to actually try to enshrine privacy into the stack instead of just ignoring it so what are some of the values of v blockchains um first of all light nodes one of the biggest kind of disappointments of rep 3 over the past decade is that everyone just uses centralized end points like inferior Alchemy but not but with light nodes especially like using for proofs and sampling we now have the opportunity to actually make it possible so that wallets connect directly to the pp network instead of using some centralized RPC endpoint we also have the opportunity to build a more positive sum ecosystem that's kind of like a evolves past maximalism and have modularis where different components can interact with each other while they're trying to compete with each other and finally I think the most important reason is because we're basically kind of stuck in this endless cycle of new layer ones so every kind of every cycle every crypto cycle there's this new day ones so like 2014 we had you know EOS Cardon in previous cycle we had you know salana avalan and in this cycle we had Su Aptos and definity and they all promise to do some to to have incremental improvement over the past cycle but it's a endless cycle and it's fundamentally extractive so with the modular ecosystem we can try to escape that cycle by having components are reusable instead of having to deploy a whole new one EOS system every time we want to make it incremental Improvement and um so so summarize and some of some some of the values of what summarize the values of mod blockchain firstly users can be first class citizens of network by running light nodes secondly we can escape maximalism and have a modular positive Su ecosystem and thirdly communities can be choose to to be Sovereign if they want by having Sovereign rollups which give you the best of both Wells so thank you
Automatic transcript — names and jargon may be misspelled.