Preconfirmations on Mainnet | Murat Akdeniz - Primev

hello everyone happy to be back so let's welcome Murat adinis from prev and um this is a workshop on pre-confirmation on Main Nets so he will explain ways to hedge um with transactions so I'll leave the floor to to morat thank you AA hey everyone um sorry for the confusion there I guess I should have pre-confirmation on uh mainnet for a long time and uh about a year ago we got on this journey and um you know it didn't resonate with a lot of people first because it required a lot of understanding of meev and block building Dynamics uh but today we I'm happy to say that um we're able to achieve pre-confirmation on mainnet and we're testing them on holke but uh before we get on uh get into all of that let's go into what pre- confirmations are so uh we published a a a research piece earlier this month that established a definition for pre-confirmation and we Define them as a credible heads up before a confirmation happens and this definition applies to many things not just blockchain so when you order something from Amazon for example Amazon gives you a credible heads up before the item is delivered in your house and that credible that credibility is so powerful that you trust it and you know like you're you're you're better off not second guessing Amazon's uh heads up that you know some pack of razors is going to be at your house and um ultimately you don't care about whether USPS is delivering the Packer razors to your house or somebody else so um we want to achieve the same kind of Paradigm uh on on ethereum and blockchain systems so um we dive into how we do this uh and and we Define these four types of pre-confirmation as transactions State block and lot so what do these mean we're going to uh go into transaction um pre-confirmation for the most part in this presentation however the others are um you know state where you claim some State on ethereum um a block so you might have a block that uh might get pre-confirmation and then you might also have a slot that is way in the future that you have some rights against right so uh these are how pre-confirmation can apply in different uh States but right now we're focusing on transaction preon informations meaning you have a transaction that you want to get on chain and you receive a credible heads up uh from a group of entities or an entity about this so um I gave this kind of Amazon USPS uh example a little while ago and it really holds under this condition and um you know the the common thinking around the space for pre-confirmation was that it needs to be delivery focused because the proposer has ultimate Monopoly on what block goes on chain or what transactions go on chain everyone was thinking okay we need to enable the proposer to facilitate this somehow but we have about a million validators so who wants to kind of have the millionth person or want to do a sales cycle to have the millionth validator adopt Your solution and then include all of this complexity and kind of these solo stakers and all this world where uh people don't want to be doing this validators just want to kind of stake their eth and then call it a day so we introduced these fulfillment um you know uh specified uh entities and right now they're block Builders and relays on on ethereum network so um we show that with a through a fulfillment Consortium model and having a network between these fulfillers of pre-confirmation you can achieve similar results without validator participation and the fundamental problem today is that uh there's no known execution risk on ethereum what you do is you sender a transaction and a group of uh block Builders relays validators um you know decide at the end of the day that uh you know it lands on chain or not and you find out 15 seconds later and through this time you have no idea what your execution risk is so this fulfillment Consortium model where you you know have the different block builders on a network uh allows you to quantify your execution risk because if uh they represent let's say 20% 30% 50% of the network you can add them up and get to a level that is satisfactory for you for pre-confirmation and then you know your execution risk in this example is 20% and then you can plug that number into your um risk algorithm and hedge your uh trades or whatever else you have um how however this model gets incredibly strong once you have the um proposer commitment as well so if the proposer said hey this pre-confirmation you guys are playing I will honor this no matter what then you know you have the ultimate credibility for your pre-confirmation so you know that the block Builder will be able to deliver that pre-confirmation promise that it gave you in real time so um how do we achieve me boost style transaction pre-confirmation there's two key characteristics here one is you need a P2P network uh for fulfillment and delivery uh we don't want to deal with single block Builders we don't want to deal with centralized entities what we want to deal with as users is a network and we just send it what we want and the network replies back to us we don't want to care what this provider or what that block Bill what this proposer says we want to care what the network thinks about our transaction and our pre-confirmation the second thing is you need a different chain to play these fast games and um fast games are defined as games uh that e does not have uh by construction because you know it has 12C block times so if you're trying to play these intra block games you cannot do them on ethereum so we're creating a chain a very kind of high throughput chain so that these kinds of pre-confirmation fast games and others that we get into later uh can be played in so when you combine these two things together you get Mev commit Mev commit is a specialized uh Network that is uh P2P oriented it's real time oriented and it has this fast chain that can settle these bids and commitments and um this is the first time such a chain or concept has introduced to ethereum so we're very excited about what that can bring now we have two things as I said um but one caveat is that it's an encrypted PDP Network because you don't want to be leaking information about what is being pre-confirmation Dynamics when it comes to Mev uh so uh it has privacy built into this P2P and then it also has this fast games chain as I just mentioned um what's different about the meev commit node compared to traditional uh you know blockchain nodes is that it's different per your role in the network so if you were bitter and you're bidding for pre-confirmation you run a bidder node and that's all you deal with and if you're a provider a provider of execution Services you run a provider node and then that might hook into your other infrastructure and your decisioning engines Etc where you can uh issue these credible commitments in real time um what's interesting about how we've set up this network is that it doesn't care whether you're proposed sequencer or block Builder or whoever else it's it's that you're on a bitter role for someone else in the Mev Pipeline and you're a provider for someone else in the Mev Pipeline and this allows for a host of applications to be developed on this platform um because basically we cannot even imagine uh this design space for and what it means for these actors and the fact that they can coordinate over this network and Achieve things on ethereum makes it super powerful so we basically think think pre-confirmation are a new transaction type uh we think it's represents net New Order flow for Block Builders and we think it's just simply new and and different than other transactions and I'm going to go into why so we came up with this uh piece of research about a year ago as you can see the last time it was changed was 7 months ago and uh I had this thesis that EIP 1559 with its gas expression of Base fee and priority fee was not sufficient to really Express what you want from the network and I had this inside and it was quite controversial uh I said hey look you need something better or you need an improvement on the IP 1559 and um I got the chance to talk about this with Tim Ruff Garden who was you know um the the game theory professor at Columbia University uh who was uh you know partially authored or if not fully authored EIP 1559 uh with barnab uh from the EF and uh to my surprise he agreed he agreed that 1559 had limitations and that such a pre-confirmation bid could off augment uh the characteristics of EIP 1559 so what does that look like um you know EIP 1559 has these two Fields Max fee and priority fee and none of them are concerned with time or pre-confirmation what we now add over Mev commit is a third field about your pre-confirmation bid and uh as you can see this is expressed uh through a transaction hash reference a bid amount in uh way so in gas and eth and a a Target block number that you want this pre-confirmation to happen against and uh once you issue this to the network in real time providers are able to give you credible commitments uh that um you know you you can use to uh understand your execution risk and take action um and this concept of Decay becomes very very important uh some of you might recognize this character which is a goblin Banker uh from World of Warcraft and uh the famous quote is that time is many friend right so in the case of pre-confirmation time is actually money the amount you bid has to do with the time that you bid bed at and it decays in real time such that all providers are incented in real time to commit to your transaction because if you don't have such a Decay mechanism what ends up being is you everybody waits till the end of the block and then pre-confirmation and it sets uh the precom bid apart from EIP 1559 or other gas related bids now how does this work well you send your pre-confirmation bid and you get real-time commitments some magic happens afterwards and then you profit well um let's let's dive into this a little bit right so what does that really mean uh well once you receive these commitments uh the actor that committed to you the provider is working to figure to how to deliver this commitment to you right and and they're going to make some decisions to do that and ultimately opt to include you in their block because otherwise uh they will get slashed so um that's a whole bunch of kind of magic that happens in between where you know your intent or whatever is being solved uh right but at the end of the day um you know you should get your result most of the time and in the case where you don't uh the provider would be slashed and you would get rewarded funds so um but in the golden case if all goes well your transaction lands on chain you knew about it ahead of time and therefore your pre-confirmation bid goes to the provider that provided you this service now we take a different route when it comes to EIP 4844 and this is a very new phenomenon we have research in this um in this space that is very very important now we identify blobs as the Rio character from Street Fighter here because Rio has a special power and I will reveal what that is to you but EIP 4844 is fundamentally different because it introduces another transaction type these blob transactions and these transactions are one very very cheap for rollups to submit their state commitments and two are very very large they're equivalent to about the largest transaction on ethereum so um it goes to I mean it might raise some questions about well why should I add this really cheap transaction that's going to like be very very big and be Troublesome and and that is an uncomfortable truth today now what becomes very important when it comes to EIP 4844 is the block latency each uh blob transaction adds a tremendous amount of latency to the block because of the size that it brings on and we go into this in in some charts now this demonstrates um how uh basically blobs might travel to be verified by proposers and as you can see they might um you know arrive at different times and that might be that might be problematic because it can't be verified in time and then it could lead to a reorg or missed slot so what you really want is to have the knowledge ahead of time of what blobs you need to have in your block and pre-confirmation become very important for that because if a block Builder can tell you hey these are the blobs that you're going to need ahead of time then they don't have to wait for these blobs to travel over the network or anything they can just send you the block and you have the blobs boom propose it to the network everybody wins so we demonstrate how sizable uh blobs are compared to the rest of the block as you can see um 90 or so uh perc of the time um you know or actually they take about 90 uh or so% of the block uh in 10,000 plus block counts that we observed on the whole ski Network so um you know they take up a lot of the block basically this is a similar chart uh it shows you what the size difference of a block is uh with or without blobs and as you can see um you know many uh blocks have um lots and lots of blobs and they add a tremendous amount of size and kilobytes to the block and this slows it down uh for propagating across the network and what's very interesting about EIP 4844 is that there is no direct priority fee mechanism so can't actually tip the block Builder hey here's some extra money please include my blob uh what ends up happening is that the priority fee per gas that EIP 1559 refers to gets stuck inside the type two transaction referring to the type two transaction details and this new field Max fee for BL per blob gas ends up referring to what the blob is but it works more like the base fee rather than the priority fee going to the network over the block Builder now you can work some fancy stuff and and figure out some solutions here but really the availability of pre-com bids makes this easy just tip the block Builder and there you go it's pre-confirmation access to this cheap uh data in this field but guess what in times of high Mev they're still going to have to pay the same prices so in order to basically abstract this concept we use pre-confirmation so that these timing games related to blobs don't have to be played KO we announced M commit on hsky test net this is available today um it's working uh we're very excited actually uh we had a major release go out today that that stabilizes the network as well and um we're on our way to mainnet to have pre-confirmation happen on ethereum um this is our Explorer um you know it it just reset because of the new release but as you can see there's a few folks on there testing pre-confirmation um and actually the block Builder that integrated is is uh I believe the dominant block bilder on Hol key test net today um I encourage the audience to check out Mev commit and um you know let us know your thoughts uh to summarize mevc commit is this new Fast game chain uh and a peer-to-peer Network for pre-confirmation on ethereum and many other fast games to come uh and we believe this is a groundbreaking moment for ethereum to uh present user experiences that are expected in the web two World such as real-time user experiences and we believe it also unlocks a new design space for real-time applications to be developed on ethereum uh this is a QR code that you can use to get more information um it has a bunch of links uh so if you scan that um you know you can read more about Mev commit reach out to us follow us on Twitter reach out to me Etc uh we're working with block Builders validators and relays right now uh Searchers are already sending bids on the network but there's a lot of room for all sorts of folks to come wallets RPC providers other protocols to um make use of pre-confirmation again this is uh live today on Hol Sky test net and will be available on mainnet hopefully by um Devcon uh in th Thailand in November uh so so um hopefully way before that but yeah um I believe we have a few minutes for questions so I'd love to hear the audience's questions if you guys have any um but yeah that's us at Prime and this is M commit thank [Applause] you hello um I would like to know how meth commit would help help to prevent sandwich attacks and stuff like that does it help at all or or not basically yeah um great question and thank you for it so we have a spicy take when it comes to Mev protection we don't believe you can be protected from meev we believe it's a fundamental fact of um distributed systems and and current blockchains so um and we demonstrate this in our reordering slippage uh research if you want to read it it was published on frontier. Tech uh where we show that even Protection Services experience lots and lots of slippage up to 10% of times um so uh what Mev commit allows you to do is fight fire with fire which is why it's fundamentally different it is not protecting you from Mev but it's giving you the superpowers that you need to beat the other Mev uh and you can use pre-confirmation for that purpose right so if you if you're working against some State and the block biller pre-confirmation guess what uh you've won that and then uh we're working on other sorts of commitments where you can specify things like hey I want to be the top of the block hey I want to be in the top 10% of the block right so you can statistically avoid lots and lots of front running in these ways um now uh is it a foolproof 100% guaranteed mechanism no it is a social commitment at the end of the day but it should hold 99 plus per of the time and that's way better than where we are today and that would be real wins and not just some fake protection that you're sold and sown some nice numbers but then you find that that you experience 10% slipage on chain uh hi I was curious how do you deal with uh subsequent latencies on the um uh on the parallel chain that you're using for um for processing the uh transactions uh I saw the 0.1 Mill 0.1 second block times and I think it's a particular challenge to keep track of time with that granularity in a distributed system thank you yeah wonderful question um so that one uh 0.1 second uh block time right now is basically possible because of its a POA chain that's um you know built for high throughput it is problematic when you try to decentralize that more and more unless you take some Robin round robin kind of approach right uh so uh we have a Innovation here uh where we're we're working on adding a Tim stamp to the bid uh and we have an a test on the network work that um attests to the time stamp that the commitment happened so the Delta between those time stamps ends up being how you want to calculate the Decay over uh the block seconds so basically having this abstraction mechanism for time allows you to um B increase your block times and still be compatible with a fast games uh mentality um at the end of the day these are signatures happening over a peer-to-peer Network and yes they do settle in a chain uh but the settlement uh is not kind of in the critical path for you to understand the knowledge about your pre-confirmation right so um because of that mechanism and and then the Decay we're able to kind of abstract things away from the um block times eventually so we would anticipate the block time to go a little over uh 0.

1 uh seconds uh but settle hopefully uh you know somewhere around 0. 2 or. 3 something below uh 0.5 at the end of the day but great question thank you thank you morat can we show some love to morat please it's been great learning more about prec confirmations and I'm very excited about what you're building um okay so we have a break five minutes and then we'll be back with a a panel