Ethereum Pectra Upgrade — ​Lazar Travica | route3

hello everyone I have a bit of a cracked voice but it won't be that yes yeah but at least I'll be in the frame so uh what I want to talk a bit uh uh shortly today is the ethereum pectra upgrade uh it's a fresh topic it's uh due to be live uh in the next month they've promised we know that uh ethereum people tend to uh uh uh you know follow on their scheduling promises quite well so let's not uh let's hope for the best this time uh the way we will approach it is that uh we'll do a quick refresher in case of somebody here not knowing how today Network upgrades in the ethereum ecosystem operate uh then we'll focus on the updates to the evm updates to the consensus and when it's all due uh it was very late yesterday evening when I Was preparing this presentation I had a killer joke to make but right now I have just the slides and I've completely forgotten the joke so it goes something like this the future might look Bleak but don't worry we are quite possibly living in the worst possible timeline and it can only get more entertaining from here so let's talk uh briefly about the past ethereum started off 20 15 Frontier Fork it enabled the Genesis block the block started producing then we got Hamstead we got the first replay protection tendin whistle evm upgrades dos protections Dow hacks happened somewhere here we determined its best to just rewrite the history of what happened and give the foundation back back its funds uh then we got Byzantium Constantinople the network kept upgrading in a pretty linear manner but what happened uh uh uh sometime I think around 201 uh 18 or 19 a completely different network started operating which we today call The Beacon chain it was completely separate from the ethereum network but it kept getting its upgrades on its own uh the ethereum as we know it continued upgrading we got the Berlin Fork the London fork with better gas incentives and monetary policy of the network and the first upgrades to this completely separate Network started happening as well in the end with uh uh the Paris uh fork in ethereum a specific event took place which we called the merge when these two networks started operating together borrowing the security from the beacon chain where today we have somewhere near 900,000 validators uh operating with the Paris version of the evm uh uh virtual machine so what happened after that is we got chappella Network upgrades start started being called as an amalgamation of these two verbs uh the best way to I I just figured out while while preparing this presentation is that execution layer or the evm upgrades are always called uh uh uh borrow their names from cities around the world and these are all stars I didn't know but they're either stars or constellations so yeah yesterday I learned um then we got Denon and the thing that we're going to be talking about today uh is the pectra upgrade the Prague version of the evm and the Electra version of The Beacon chain so first things first uh Electra handles everything when it comes to security validator at the stations votes entering the Network as a validator exiting the network as a validator receiving rewards uh whereas Prague just takes care of our transactions being executed correctly and these two uh uh recently have a new uh uh have a new thing where uh uh a new intertwin dance together which are the blobs which are ethereum's answer to uh cheaper data availability so let's first talk about what are the changes to the virtual machine what are the changes to the execution layer that I think are quite notable so uh uh it's a pretty unimpressive upgrade when it comes to uh when it comes to the evm changes we've done some quality of live improvements it's easier now to configure your execution layer nodes and to tell them uh hey this is how much blob uh space we are aiming for the blob schedule is written in the E configuration no more complex communication between your consensus clients and execution clients uh uh general purpose El requests also aim to simplify how these two nodes operate why I'm talking about this because it's important to right now understand the trend of people soon soon we are talking in the next 3 years uh being able to completely separately run these two uh uh uh networks statelessness and light clients might help this but it's uh pretty clear that with the eips that are being scheduled people are less and less focusing on heavy communication between the CL and the elel locally of course uh we're increasing the cold data cost something that uh we can now do without worrying that rollups will start abandoning the ecosystem simply because we now have blobs something around 70 to 80% of all rollups have completely upgraded to using blobs uh uh uh uh for uh intermediate storage uh we have a more secure BLS curve so signature aggregations no longer depend on 80 bit uh uh key sizes right now with uh this uh specific curve we can go above 120 bits uh one more important thing uh for people who are designing bridges for people who are designing oracles uh which are completely which are taking up a lot of gas on the ethereum L1 uh we will have a system smart contract which will uh store a recent history of block hashes so oracles will no longer need to operate their own source of Truth when it comes to block hashes for people to uh uh submit uh witness proofs to but it will be tackled on the protocol layer itself but the thing that I think is uh uh most important and it's a bit confusing when I first heard uh that it was being scheduled for the next upgrade is the next uh big thing in account abstraction which is the eoa account code or EIP 7702 so that's what I want to uh talk a bit more about so in essence what eoa uh uh account codes mean is that you can uh temporarily or ephemerally set a specific authentication code uh set the pointer to a specific authentication code to your externally owned address that means that your let's say metamask address you can instruct it temporarily to authenticate using code from already deployed smart contracts you would basically say hey when I want to interact with this EA use the uh uh use the bite code that's deployed in these smart contracts and this can be anything you want for a complete Suite of uh what's possible I instruct you to take a look at uh uh uh safe contracts which have a lot of plugins for authentication uh which support web auen which support all these different curves which support uh all of that but they rely on somebody else bundling the transaction and sending it for you this being a protocol level change you would be able to do all of this on your own from a wallet that starts implementing and supporting 7702 right now we have privy who's doing that have you heard of privy o yeah some of the people have uh uh they are they were pretty big uh and pretty involved with this EIP uh but on the other side we have no clear indication that other wallets the the biggest wallets of of them right now on the market are really following this direction hopefully they will we will talk a bit about how this changed so the next thing is Dex gas payments hey I'm a new user everybody when they're talking about account abstraction they're thinking about new users how do we onboard the next billion people to using ethereum well it's possible to do this as with any AA uh uh scenario it's possible to swap your Dodge Coin for the native currency pay the gas fees in the native currency and execute the transaction that you intended and uh third which is for me uh uh most exciting if if you were if you were ever swapping anything on let's say Unis swap you know how how uh irritating it it was how still a bit iritating it is that you have to click on multiple prompts to send first your rc20 approvals and then to initiate the swap uh with the contract itself well we can have Atomic batched transactions without added infrastructure and uh uh uh smart contract gas fees uh next uh and uh this is the part that uh that I I presume will be pretty wild especially with all the agents uh as the right as the current hype uh uh in The Wider um industry AI agents as the hype in The Wider industry uh you will be able to uh uh set temporarily uh authentication uh mechanisms on your eoa and a completely autonomous let's let's say it's another EA or it's an invoked smart contract it would be able to uh execute uh things on your behalf meaning that for all solidity de for all smart contract developers here meaning that the origin of the transaction as seen by the origin op code would still be your eoa so this is pretty big a lot of a lot of uh you know incentivization mechanisms a lot of uh uh uh careful thinking has been done uh uh uh uh so far to circumvent that we don't have it uh built into the protocol but this uh is huge for having session keys for enabling uh less friction in web free gaming uh and for uh doing all sorts of stuff with autonomous handling of your funds that you don't have to click anytime your AI agent wants to do something I'm not that uh uh acquainted with the use cases for AI agents uh in crypto so yeah forgive me but uh the the the topic that uh somehow begs us to talk a bit more about it is what two years ago can somebody correct me uh year and a half ago we had ERC 4337 as a pretty gentle introduction of account abstraction to ethereum I myself don't see a lot of uh a lot of benefits in the software that I use every day to interact with the chain that 4337 brought me uh but it's idea uh which can be seen by the prefix itself this one is an EIP this one is an ERC what's an ERC that we all know it's ERC 20 but when people first thought of hey let's have a standardized uh uh uh let's have a token standard that all wallets can interrupt with seamlessly it was a request for comments ethereum request for comments people didn't need to change the protocol to make it work people were just discussing about standardizing certain things such as IRC 721 for nfts such as everything else that comes with an IRC in its name so ERC 4337 the sole fact that it's an ERC that it doesn't change the protocol of ethereum makes it gentle as I said uh when I was talking about it and what makes it gentle is that all the introductions all the uh Technical Solutions for enabling us to bundle transactions to uh include uh sponsored gas transactions on somebody else's behalf that is all done pretty much off chain you have an alternative mempool that alternative meol meol is not a part of the network itself as we experience that every day with me um but this was a pretty gentle solution in uh uh in a sense where you to consume the account abstraction uh uh offered by 4337 you had to integrate with basically a third party Solutions provider you had to find somebody who would bundle your transactions for you who would verify who would help with verification of your transactions and submitting them on chain so what EIP 7702 does it goes headon to to implementing account it plays well these two play well but what I think uh why I think that 7702 will have a much bigger impact on the ux is because we're not solely thinking about hey how do we bring the next billion users on chain without messing with the Protocol no we are thinking about how do we do all of these validated things that people are doing with save contracts for example where they have complex authentication schemes a complex ACLS however you want to call call them how do we make them uh uh a part of the protocol where we don't have to rely on good people from safe in maintaining and auditing their uh their code we don't have to rely on the bundlers that they provide good uh uh uh bundling relays for the transactions but we have ethereum providing for us what we need and that is custom authentication methods gas sponsorships and even as we saw here autonomous transaction executions even on eoas uh I'm going to pause here for a bit yes this is the end of the machine part does anybody have any questions about it yes please engineer do you think this new e 7702 has the potential to uh impact certain companies protocols that are building on 4337 could it remove some of their use cases yes definitely definitely uh in a similar sense as Pier Das would in my opinion and I'm always going for let's have it all in the protocol if you validated that uh what you as an infra company in web 3 are offering to the people your biggest threat is not your competitor your biggest threat is the Smart protocol guys from the core team implementing all of that in the protocol ER so both blobs and this are a huge threat to to some shitloads of VC money in this sphere yeah uh any other questions about no okay uh let's move to the consensus part so this is the Electra Electra is a star as I said very important to remember um we have blob throughput increase people thought that we were pretty conservative about the limits that we set for blobs uh and it artificially Rose the prices but we see that we can increase the blobs without sacrificing uh uh the BL throughput without sacrificing decentralization uh Community index outside uh attestations uh this one is related to light llies uh actually we would be able to with this we would be able to uh more easily create zero knowledge proofs of ethereum's consensus operating correctly which is huge for light clients um and uh uh validator deposits uh on chain this is a let's say uh a quality of life Improvement so not that important this one is important but I hope we'll have a chance to talk about it once we have uh true statelessness what we have once we have true ability uh to instantly instantly uh to to verify the the correctness of the blockchain without uh uh having to trust anyone or without having to completely execute everything on our own uh what I want to focus and what really uh got me excited about this talk is uh these two eips uh they might sound uh pretty simple but they were not simple and even they took a lot of uh think careful thinking considerations uh uh uh how it will impact the the network incentives well the max effective balance incrementation means that you still need 32 ethers to become uh uh a validator in the network but you can stake up to 248 what this means is that ethereum uh which I must say I love ethereum because of this nobody thought hey but if we increase the limit to 2048 these these people and companies with huge bags of money who are staking their e and who have let's say 200 validators that means 200 time 32 ether locked in a in a single uh um validator Hardware validator on the Chain but acting as 200 of them that means that our statistics when it comes to decentralization will appear less well no it would they will appear more fair and we would finally be able to know how much valid how much validators there there really are in the network nobody nobody would have the incentive of further registering new validators for each 32 ether that they have simply because of gas cost simply because of operational costs and a bit of bit more of Hardware resource usage so if you have somewhere between 32 and 248 ethers you can freely stake that amount and you will be rewarded uh appropriately so this is a big thing in my book and we can expect uh because there are migration uh uh paths for all of these uh there's a I forgot the name it's not consolidation but it's something like that there's a path for people to uh apply for joining their multiple validators together and having simpler staking operations on their end so yeah we can do that now and the next thing is uh you've all heard of I don't know Lio figment companies like that that offer non-custodial staking Solutions well uh it's not that uh uh it's not a well-known fact the gaping hole in security and how much trust even non-custodial uh uh staking Solutions uh so far required of all the consumers of all the users of those Solutions so let's uh let's see where the problem was and how it is fixed so um as a refresher the way staking in ethereum functions right now we have two sets of keys we have a withdrawal key the this is a plain simple uh regular eoa on the Chain where you as a Staker uh you would sign a certificate saying Hey I want all my rewards to be distributed this address and this eoa always lies with you on your Hardware wallet somewhere this is the address that's receiving all the rewards that you have and this is the only address to which uh the stake can be withdrawn once you exit the network as a validator the other uh uh uh uh set of keys are the consensus keys this is a completely different curve this is a completely different use where you use these keys to attest that you have that your subnet committee or whatever hierarchical organization in the consensus has validated a certain block he has validated that the block is correct and can be included in the chain so these consensus keys in most cases are operated solely by the non-custodial staking service because we've all said hey that's completely fine uh uh I know that these people cannot steal my rewards I know that these people cannot withdraw my stake to their account the worst that can happen is I uh is I stop being a validator and I get my 32 ether back well that's not exactly the case um you as a Staker as a PE as a person who's using the non-custodial staking service uh sometimes you would want to uh uh unstake and get your 32 ether back to buy a house well it's it's the non-custodial it's the institution that's providing non-custodial staking services to allow you it's up to them to allow you to do that because they have to vote using the consensus keys that your withdrawal account gets the stake back which is a pretty broken game of incentives if you think about it uh you're getting the rewards with most of these services but what you're not getting you're not getting the entirety of the uh uh uh gas uh uh gas fees when you are the lucky winner for producing a block the majority of those funds usually get the staking services and it's the that is the playground where they compete with each other who has the less fees when it comes to proposer rewards uh in this case they don't have an Inc incentive to let you unstake we haven't had huge incidents of big companies uh uh blocking preventing people from unstaking but it's pretty clear why they would potentially think about that so luckily what uh uh uh where ethereum is going is that in a month in less than a month from now uh we will be able on the execution layer meaning uh on transactions that are popping up on ether scan you would be able with your withdrawal key uh to uh to request uh the uh uh validator exit and unstake meaning that your uh um whatever you want to do is in your hands not in the hands of the non-custodial staking operator uh which what I expect is that more serious people start entering these kind of things people that would get themselves informed about how all of these work this works uh before getting into it of course uh this broken incentive system brought us some wonderful tech such as distributed validator technology where you always have to keep some Hardware online that's much less uh uh expensive to operate in order to have the uh consensus Keys sharded with you and your staking service but right now again as engineer pointed out a lot of these upgrades are slowly making some complex and really cool Technical Solutions that we needed a couple of months ago obsolete which is how a network should evolve um yes we talked about uh the triggerable withdrawals and I'm going to end the talk uh luckily uh 6 days ago uh there was a a core protocol uh uh meeting they're all public on YouTube there's no ethereum inner circles that I know of uh where they determined uh it's too much of a risk to include evm object format it would be very hard for uh uh Lazer to prepare this presentation if they determin to put it into the fork uh so luckily they didn't uh They delayed P Das uh which was a lot less uh uh mature anyway which is the data availability sampling protocol uh and uh the Petra as I explained it today uh will go live on test Nets in a matter of weeks and hopefully in the beginning of March we can see it uh on on the main net that's all and yeah let's keep etherum awesome thank you [Applause] uh I forgot to ask are there any questions about the consensus part of the things yes please M uh the withdrawal keys will probably work from now so are they compatible with backwards comp yes yes everything is backwards compatible in the case of uh in the case of the uh El triggerable withdrawals uh there's no migration path it's a drop in upgrade no but the withdra key how does the network know for the withal keys of the current users who have with the noning provider uh because uh uh because you you are you you have a certificate on the consensus layer that says hey you cannot change my reward address this is the part uh yeah I didn't draw it but this here are the are the yeah rewards accessible by withdrawal key this is a known fact so there is a a a mapping between the two okay thank you any more questions thank you