DVT, Empowering POS Decentralization | KBW2023

hello ladies and gentlemen uh I uh my name is John Park I represent a41 uh which is a validated company based in Soul South Korea and I would like to actually present you uh obal Colin Myers who is a CEO of obal um and he's uh he's brilliant idea behind like uh DVT technology and how he's trying to bring this idea to real life uh Colleen would you like to introduce yourself to uh the Korean public hey everyone uh thank you for having us super excited to be here um this my first time in South Korea actually since before covid uh yeah so super excited to get here and eat some good food thanks for having us so uh it reminds me of the time when we first talked about the uh the idea of secret shared validator a couple years ago and it it's also very uh enlightening that we're actually on stage together to talk about this idea um could you actually uh explain to the I mean Al EAS kind of mentioned about it in the last session as well it's to say laying the gra groundwork for us but uh could you actually Define what distributed valid technology is to the Korean public and the audience here yeah absolutely uh the easiest way to explain this is uh distributed validator technology enables you to run an ethereum validator across multiple machines uh today that's not possible uh today the only way to run an ethereum validator is on an individual machine uh and over the course of the talk we'll kind of layer into these different facts but yeah most simply it enables you to run it across multiple machines which you can do a variety of different things with so um with with DVT being implemented in real life what would what would be the what be what would it look like in terms of like you know what differences would people actually notify when they're staking or like from a it could be like delved into like in a few other areas like from few angles like from node operators perspective or from a Staker perspective as well like Could you actually share yeah so the technology itself is meant to be implemented as a primitive and meant to be used as a building block um liquid staking pools as a relevant example are the most futuristic use case for the use of DVT today uh however it also has implications for atome validators centralized validators uh dii applications and even today DVT is relevant for some of these liquid staking token backed stable coins and some of these new products hitting the market so um yeah in that sense it is a true building block um happy to expand a little bit on some other parts of how it can help liquid staking pools not mentioned in the previous chat and also discuss maybe a couple of other use cases uh for different users however from where I sit a lot of the early adoption of DVT was focused on uptime uh essentially downtime prevention however now where we are at with the technology and the Primitive is it's now being used for more futuristic use cases uh which put its cryptography uh to the highest test uh and at the end of the day DVT is really an applied cryptography project uh that's being turned into software so a more specific examp example as to how it helps LSPs in that regard is U today in liquid staking pools uh all liquid staking pools no matter what it is or the size of it or the brand of it if there are multiple validators that are a part of that protocol uh each and every one of those individual validators are responsible for the machines and the private keys that the algorithm sorts the stake that ends up with them uh today in some of the larger staking pools there are operator valid ators with between 100 to $500 million of value sitting on their own personal machines uh today there is nothing preventing validators in that context from shutting off their machines uh going haywire or potentially trying to be Byzantine uh in some nature uh today all liquid staking Pools use social economics Andor social punishment or being voted out uh or some other type of gating system to prevent Mali or evil Behavior Uh however with DVT uh if all of those validators are sharing the same key set then not any one of them can act malicious or sabotage or selfs slash or exhibit downtime towards the pool uh so yeah that's that's pushing the DVT primitive more towards its Byzantine behaviors uh which has been very quite interesting for us to see it's today our primary go to market uh and where you'll see DVT utilized in its largest scale on Main net first today we are on mainnet but not in the thousands of validator scale uh the first place everyone will see DVT utilized at that scale uh will be in liquid staking pools um in addition to liquid staking pools there there are other benefits uh on the institutional front I will also touch on a little bit later we'll dive into it deeper but there's also benefit for your centralized validator uh this is your stas business your traditional stanking as a service uh kind of the backbone the industry what was built over the past four years and the company and Enterprise type that provides all the infrastructure for Lio uh for these providers it's actually proved to be very very beneficial in the cost-saving element uh it has enabled all of these different validator entities to uh decrease the cost of their machine setup while increasing the security this has been super important for helping these stas products scale uh as most of them are were early to the market but are now being heavily competed against by liquid staking pools and and having honestly the majority of market share is kind of coming out of stas and it is going into pools and to stay competitive DBT has been a way to help them decrease their cost in addition to that it has helped them substantially with their insurance topic uh which is another area later in the chat that we'll dig into yeah and uh I kind of briefly want to touch upon uh how DVT could actually bring some Regional diversification as well um I think we we could actually touch upon the idea of cluster together um and for cluster like you know I think there are various types of clusters uh you can definitely like do more explanation about it and at the same time give some illustration to the audience about how like you know we can hypothetically form like a multicluster together like you know throughout the region and how would how it would actually help out the regional diversification or diversification of nodes um which will help or cont contribute to the decentralization of eth and and many other networks that will be able to adopt DVT in the future and opal yeah happy to spend a couple of minutes on geographical uh requirements and their importance so uh today the Obel core team is 28 people spread across 14 countries uh we have teammates from San Francisco to India uh and as a result of that we've been fortunate enough to push the limits of this technology just within our own core Team all across the world uh so the the the first two validators that we put on Main net were entirely run out of people's houses and also multic continent and multicountry uh the first example we did is we had two key shares in Ireland one key share in Estonia and another key share in Canada uh that was the first distributed validator that we had put on mainnet uh and that was a way to so the reason that we chose the countries that we did there to be more specific is the one key share that is in Western Ireland land uh traditionally is an at home validator like location and uh it has very poor internet uh and and the person that normally lives there actually usually runs an at home validator uh and said look I actually think with me traveling so much that this isn't going to work uh so if we do want on Main net as oble we're going to have to share it amongst people in different countries so as Izzy alluded to earlier as well not everyone has real internet and this has been a huge Learning lesson for me is we're now running distributed validators in South Korea we're running distributed validators uh in Australia and New Zealand and India and you know most places that I thought had fine internet to run an atome validator but actually they don't um so that that's been a very interesting uh geographical redundancy element to study and find out which is the reality of it is internet's not even that great for there to be a whole world of at home validators so having technology that enables tolerance is actually literally the only way that we'll be able to scale this network uh in that regard um right now on the growth effort for obal in our adoption cycle uh next year we will be focused most on localization in global areas that we have found High adoption over the past two years uh right now South America has been very very interesting for us last year the ethereum foundation had Devcon uh in Bogota uh and then prior to that did a road to Devcon all through South America and we've now like tapped into that community of people who learned and then were onboarded into the ethereum space and there's lots of node companies node Runners projects middle markets small tier validators all coming out of South America so uh we find that as a new area for like growth and hope for the global validator Community uh and that's now how we're focusing our global mization efforts with with obal is where is the validator count low on mainnet uh why is it low how can we help uh and then ultimately pushing and promoting people to use distributed validators in a geolocation based way uh you know you can use it you can run them all the key shares in the cloud if you want through Amazon awe so you know you can also do that um or you can divide your validator up amongst four people and four continents and you can see what you get and that's the type of testing we've been doing on mainnet today ourselves and also in the testing efforts that we've been doing with Lio is like really pushing that geolocation to the Max and checking latency and saying the added benefits of latency do they are they to worse than rewards so I guess what I'm trying to say there is a distributed validator in multiple geolocations increases latency latency in an ethereum validator normally means that you get less but are there ways to rationalize security versus latency uh and that's really what we're testing for now for geolocation anding adding a little bit more um I think there are two types of uh clusters as well solo and uh multi clusters and um not just uh like you know Regional diversity that will be added but I think there should be a lot of other benefits or of running or there should be like a reason why uh like people would do solo cluster and multicluster as well what are the perks of the too and uh I think there will be a lot of a potential node operators here or people who are interested in becoming an at home validator as well so like who are ideal types for each group in cohort I think people will be interested in in hearing that yeah so um today there's a lot of eth staking products and at the beginning of this there were not um and as a background I I used to work at consensus for a long time and I used to work with the EF quite heavily on the initial adoption of the E2 spec and the Genesis event um so so we've we've spent a whole lot of time on the adoption of the different user types and and making sure that there are different user types um and to today what's most interesting is uh By Design it's worked out now that we're three years into it there are a lot of different validator types and each validator type has their own wants and needs and that has resulted in multiple different products uh and the best example would be kind of like to the question that Victor asked earlier about you know Lial vers uial uh the way I look at the two things is is they they have different user types uh so they're kind of fundamentally different products at at at that end so what we found is that the more product types and the more different type of staking products that are developed the more different use cases we find for DVT so liquid staking pools are a multi-operator setting so they're only using shared clusters uh and they enjoy using shared clusters because it increases uh the amount of Byzantine fault tolerance that you have so it decreases that risk uh as a centralized solo cluster person uh there's a lot of benefits for for me um one of them actually comes through security credentials and sock one and sock 2 compliance um I'm I'm able to never have a full private key that could be stolen or that can go Rogue or or that can be tampered with uh so there's a variety of reasons for credentialing um that like some of the larger validators are using it and again for insurance costs um there's also a very interesting thing that we've now seen as the obal community which is um a variety of small to mediumsized validator entities are now spending a lot of time testing obal as a means to help credential lie them in other ecosystems uh which I thought has been pretty interesting to observe and from where we sit on that topic which is a new one for us is uh the majority of the large liquid sinking pools have funded uh efforts with external researchers and external parties to figure out how to objectively uh decentralize the validator set and I I got news for you it's super hard problem and we've realized that we can participate inside of that as obal by enabling people who go through our testing experiences our documentation and our community to help credential them so that liquid staking pools have a neutral way to say like this validator is experienc not just because this validator ran in a lios sponsored test net or a stake wise sponsor testnet right or any of the other liquid staking pools so we it's helped us build our credible neutrality as a protocol um we as obal want to be used by everyone you can think of us as a layer two in that regard uh we're not here for any one specific User Group we're more or less right now on a journey helping each and every new user group as they get created that are focused on building staking products indeed I can testify as a small to midsize validator who was growing at the time when we first started like joining joing obal I think it really helped out in terms of us getting recognized by all the other node operators who are on the on the obal together and at the same time like you know when we were entering other ecosystem it always helped out too like uh Lio obal test net pilot operator was actually a very joyful experience because we had a lot of conversation with other operators uh but at the same time through that we could actually build a reputation as well and I think like for any potential like you know Builders who are trying to like you know jump into the not operator Market could actually think that way um and and and and uh jump into OBO like ecosystem as well and Community as well um next I want to kind of like you know uh touch upon the the uh the the the like the road map of obal cuz uh it started out so when when it first started out it was just a paper right the idea that vitalic was actually throwing out uh under the the the term called secret shared validator and now like you know it's it has upgraded it itself to distributed Val technology and OBO is one of the one of the projects that is actually spearheading this efforts and a lot of the node operators in the E ecosystem knows about uh the Obel and what's the road map that uh you have in mind for uh Obel and is way ahead in terms of 2024 yeah that's a great question uh today obal is on it's road map to V1 and I'd say we're about 80% there uh we suspect to enter V1 territory sometime early at the beginning of next year uh and our V1 Journey so far has been about two years however DVT as a topic has been being researched and thought about in progress for nearly four uh and it has taken us this long to get to V1 uh however I'm confident that our V2 V3 Innovation Cycles will be a bit more more as law-based and will be a bit quicker um so so today we have two parallel work streams going on inside of obal one of which is the V one effort uh this includes a collection of validators that are already on mainnet it actually includes more than 40 entities that are participating in mainnet in some sort of shared cluster State uh and that's kind of where we'll go to market is with the V1 and that's how people will start using interacting with obal at scale is is with a more mature V1 a couple of items on the to-do list before we get there is really just auditing and hardening uh there's not a whole lot of feature requests or anything in the road map that is too complex or tooo difficult to build pretty straightforward however over the past 6 months we have begun to expand our research work streams and focus on more of what we're calling Obel V2 and everyone can think of that as ZK DVT uh yes we are throwing the ZK on top of the DVT um because it should uh not because it's cool but there's a variety of very cool tail-end research problems uh a specific problem called proof of participation uh which is basically what happens in a shared validator cluster when someone is being lazy uh how can you punish them or even objectively understand what their behavior is uh and that involves some very difficult cryptography and is more of an industry problem it's essentially a cryptographic problem where it's uh difficult to objectively identify in a threshold signature setup who's responsible for what uh so we have partnered with nethermind for that effort and we have also incentivized nethermind to become kind of a core development team for the obal network uh which takes us to uh the end state of our V2 Vision which is multic client so um I've spent a lot of time at like the core protocol layer with the foundation and I am a big proponent of the multic client model I I believe it is ethereum's probably greatest pillar to success and ability to stay credibly neutral is the multi-client model so for v2 our our main goal is to launch multic client improve the level of cryptography that's associated with a specific cluster uh to to make it more provable uh which is very important in today's world making things provable and then ultimately we we're already have begun to focus as well on what's called an obal V3 uh there's no formal uh effort really put together here but opal V3 is when this becomes a network uh and inside of what becomes Network today what we're most valuable for is people building staking products come to us and say who are the validators and a year and a half ago that was pretty easy because there was only like 25 validator entities uh today I think we're probably getting somewhere near a hundred or so uh and it's been proven uh with the community and our users that uh we're a helpful filter of sorts um and then now we're kind of starting to see very organically at the community level level the V3 come together uh which is really more of a coordination based effort where people plug directly into obal to build on top of uh and that's when we can become most like a lay or two in people's eyes is once we push forward with more of a network model which is currently V3 thank you um and I didn't know you had V3 coming but I was only thinking of V2 are you uh just briefly like you know before we go into the next question are you open to so I I know the nethermind is working on V2 together with obal team but are you open to actually like opening up a research or Grant towards the community for the V2 like you know ZK area that uh research that you want you want to do yeah that's a great question as well um we have started to roll out more formal working groups one of the things that obal is very lucky to have is you know we're not short on smart people who want to help and it's taken about two years to mature the project to a coordination level and tooling level where we can actually go out and work with you know people like Dan Bonet from Stanford to help progress V2 right and we can actually begin to include some of the smartest people in the world who have raised their hand and said this is a huge problem we want to help with that and now it's up to us to facilitate that and make sure that the smartest Minds come together so we've been working in small working groups that are not necessarily very formalized or public and now as the protocol mature we are moving things over to the community more and it's uh really because we're fortunate to have one uh like a very strong group of people who are willing to help out that we don't have to incentivize so you know obal doesn't have a token we're we're not a token project so I'm sure some people can imagine it's quite difficult to incentivize certain behavior um but really what happens from my experience in this is the only people that end up working with you are people who actually care or that can help uh and if you're designing from that basis then it's actually the best type of community help to get uh however you know we are fortunate in that regard not every project is able to get very smart people to help without uh having to incentivize them in you know a certain way true true touche um as a as a node operator uh when when I look at obal uh we've been participating starting from the early stage early early phases and now um we're one of the solo clusters on Main alha as well the experience so far has been has been great but as a node operator I think there are like you know clear incentives to actually pay attention to Obel but what do you think um why why do you think that investors institutional investors even have to like pay attention to DVT and obal like in a sense like you know what would actually benefit them to uh to in terms of like you know paying attention to obal and what do they get out of it great question as well so if you believe in ethereum then you have to believe in obal it's just really the only thing um DVT as a primitive was also designed and created by the foundation uh and kind of released to the community and has been specified in a certain way to enable it to grow and to flourish um so the way that we look at this is quite simple which is if if you believe in ethereum and you the network to succeed then you have to want DVT to be adopted because it enables more machines to come online it prevents single operator failure it decreases risk it increases the security of everyone's setup it increases the credibility of the ethereum network um so you know if you're not long eth then you're probably not long oal but if you believe in ethereum then you should definitely believe in DVT yeah and uh I think institutional investors would also benefit out from using like like obal uh operators because they will be paying less in terms of fee because I would probably assume that operators who are running on obal will probably have to will pay less in terms of insurance when it comes down to the uh when it comes down to like insurers like they will be asking are you like running your uh nodes on on top of obal or not and then they will probably like cuz cuz as you said like you know uh it's not just about downtime but also like you know it it just prevents a lot lot of the risks uh factors that are related to the uh node operations and for eth is actually the the threat is real for which uh operators you choose so like could you cover a little bit on on that end yeah so today well the way that the community has presented DVT uh shout out to the liquid staking pools and everyone for helping us push the narrative forward but it has been definitely presented in a very Rosy manner uh it seems to kind of help solve all your problems s you know the the end all be all whatever it may be uh but there are inherent risks with adopting new software and introducing middlewares which is what obal is does also increase complexity for the user today what we're most focused on in ethereum is enabling more validators to come online and that's not just by you know enabling DVT to add more professionals uh it also needs to add nonprofessionals into the network which is very important um so deising the network and enabling these people to to come in is super crucial um more specifically in inside of the insurance realm is where this comes in for the institutions uh we've gotten a lot of encouragement that Obel does decrease risk which has been the big litmus test that everyone wants to see from us now is you know how does it decrease risk uh please show us that you know some empirical evidence um throughout all of our different test testings we've recently been trying to slash ourselves we've been trying to figure out how it could be bad or Worse performing in our most recent larger Enterprise testing effort We performed within 1% of a regular validator uh and at this point we are pushing a threat model publicly to show people how obal could be attacked and why uh and most of these efforts are coming from insurers uh to kind of round out the thought so right now one of the more institutionally successful things that's happened to the Primitive has been an acceptance of the Primitive by insurers and now insurance providers have come to Obel and said we would love to ensure staking providers however uh today with their current technical setup minus DVT a bad day at the office means you could lose 32 eth or everything and that that is the case today in the most serious flashable situation you will lose all 32 of your eth but it's very hard to predict due to correlated slashing and kind of all these network-based uh things that no actuaries have have actually really ever seen before um so receiving this feedback from insurance has been very very helpful to understanding what they they want to see and the different adoption that we need to push forward um and we're looking forward to working on the risk and what types of risks it increase and what types of risk it decrease uh however today netn net uh DVT is being presented by the insurers as a discount uh because they are currently unable to underwrite a ethereum inv validator that could lose everything in one day yeah I mean for institutional investors who are trying to stake their eth or trying to have crypto exposure and staking uh as one of their strategy later on should definitely take into account of like you know when they're staking whether the operators are using DVT or not right um and uh just like Victor did I would like to do it too um if like uh the audience wants to learn more about Obel what are the great resources for them to actually check is it your website or is it the Forum like I think there are multiple channels that you can actually share with the audience yeah so best place to follow us is ob. TE um the most upto-date place through ob. Tech is our doc site our doc site's actually Linked In directly with our codebase so every time we push updates or or anything around that realm it'll go through the doc so the docs are the greatest source of Truth uh we also have a Discord community that you can join uh we also have a research form uh and all of these links can be found uh on the ob. tech website as far as I know Obel has uh many Korean operators and also uh is trying to achieve big things in Korea in the future as well um so if there is any final remarks that you want to give to the Korean audience I think this is the chance yeah absolutely yeah very excited about the growth opportunity of South Korea would love to see the validator presence kick up here um shout out to all the South Korean validators in the room and the investors that we have um we appreciate your guys support and super critical place to get machines online and looking forward to helping um I'm here with my co-founder and also our chief growth officer uh we'll be able to answer any questions that you guys have and uh looking forward to it all right thank you so much thank you so much for listening ladies and gentlemen thanks [Applause] [Music] everyone