Double-Sided Decentralized Validator Pool: Pooling ETH and Hardware for Staking - Eduardo Antuña

hello everyone i'm happy to be here today um after this pandemic see again people face to face is incredible the buy from this conference is really amazing and thank you for being here for for my talk um today i'm going to try to explain one of the big challenges that we have right now in ethiopia since uh we want to have a decentralized political pool but taking care of the both sides the hardware and also the the pulling the heat and yeah well my name is edward antonia i'm one of the co-founders of domino we started this project in 2017 with the help of geordie valley and also the frame at that moment we realized that even we are creating amazing decentralized technologies we are we were still committing the same mistakes that in the past so we are we were active and we are um in a centralized manner so that is why we create this uh concept of tamnot with idea of create a public good that allows to create a human-centric decentralized cargo infrastructure so if we take a look to the f2 designs goals we can see that we have travelage points come of common with that decentralization part and also the resilience but but also also the security and simplicity where um component common points with um uh with dabnot so that is why we um uh introduce east to validation on danov from the very beginning at the genesis moment um right now we are very proud that we have 2609 validators is for us an incredible number we are very proud of of having 251 million total value on damn not but uh we we only represent less than one percent of the network um we don't have metrics it's one of also of one of our principles we take care of the private privacy of the users but this is reported with the graffiti on on the network but the question is is easy to really decentralize and if we take a look to the numbers we can see that bynan has 24 000 validators on the network dragon has 30 000 validators and coinbase has uh 40 48 000 validators uh so if we take a look to all the numbers we can see that the 10 entities on the network control 62 percent of the network um we don't think that is really the centralization and that is why we are trying to achieve and find a solution for this um there is so there is also people in the community that they realize that this is not the ideal situation like for example marty koppelman in may also complains about this we haven't achieved any improvement but also for example shasha is complaining that we are not taking any measure to achieve a more decentralized network we know that when you are centralized you have a lot of risk for example we have been seeing rook pools atlassians and we are also losing all the benefits of decentralized technologies like privacy and resilience so [Music] we are also same we are not perfect and [Music] it's not just about validation decentralization we also need to talk about client diversity we only support prism right now and we are being we have been working on on be able to have all the clients from one year ago more or less um right now thank you to the ap 3030 and also the key management api uh proposed by dub lion we are very close to have all the clients from damn not thank you to the work of pablo mendes and also marcelino we are going to revert this situation um this is our current model um we only support one client for one one reason is that our aim our target is to make as easy as possible for users to run a validator so we expect then not to run command line the only client that introduced our ui was prism and that is why since then we have been using but um we need to change this and for that we are going to start using web3 designer we are very close to to have it ready we are in the final test phase and the good thing about the web resigner is that we can control all the clients with just one piece that aggregates all all the keys of your node and this is going to allow us to switch between clients and it has it has also several properties like for example having one only slate slashing database you can safely move from clients and when you target people is people that are not technical this is a critical it's a critical point now with this solution we are safe that no one is going to make any mistake to make this happen we had the help of the president people like james also joaquim helped us to build the western center ui with which we are going to control all all the clients but also all the help of several client implementers so yeah let's talk about other themes let's see why it's not possible today to have a truly decentralized validation and a key point of this is that you cannot delegate your validation key to anyone because he can slash you sign an exit or research you so that means that you need to control the validation key and you can't can put in the hands of other people so this is the main reason why people need to set up their own hardware and infrastructure and to address this situation we figured out how we put how we can put these four pieces together vls signature has incredible properties uh these properties allow us to create a three echo signature uh nf of m that we can use to create a public common public common public validator key so my idea is that if we sign the same message with different private keys the aggregation of nfm allow us to have a common public public key but also we have another interesting thing that is the distributed key generation this multi-party computation allow us to create a private key well not a private a public common key without having the private key at any moment on any in any place this is quite different to the same secret same secret sharing because as you may know when you have the submission you need to split the private key so that means that you need to have the private key at some point with the distributed key generation you don't need to have this kit at any moment and also we are going to take this with three any that are already mentioned and making some kind of magic or well more cryptography that's magic we are going to we are working on this approach so the main idea behind of this is to take a damn knot as is right now and add a new piece that is the web3 aggregator it's quite similar to the web3 designer but with slightly difference so we are going to take the duties of the network the proposals the blogs and all the stuff that need to be signed and in this case we are not going to sign it directly like in the west resigner we are going to broadcast to a consortium of the dkg signers so we want to split the validation work from designer work when we broadcast these messages to all these entities then we are going to receive the response on on the damn node and inside of that node is where we really aggregate all these signatures this has several interesting properties and the first one is that since we have a threshold signature and of m the n first response responses are the one that we are going to aggregate making the latency latency to have an asymmetr very very small that is very important when you are validating on estrogen tube and when we aggregate this signature then we have the silent message that we are going to broadcast to the network another property that this model has is that we can have management of these uh validators and also designers allowing us to opt in opt out of the network that currently is one of the issues of steering to you can sign an exit uh that means that you are stopping to validating but you cannot opt in for one month or two months so imagine that you are moving to a new house then you need to sign an exit and when uh wait until you can take the ether but with this you we can have protocol that coordinate who is the current validator and distribute the network in a more uniform way um yeah well mainly this is a sum up of all the benefits of this is is trustless since we don't have the key the private key at any moment increase the security since we have we can have slicing database on the hardware but also on the on the consortium of signers um it's it's quite simple to implement quite not easy but um at least we don't need to modify any of the current client implementations allowing us to have a plug-and-play solution on top of them we increase the centralization i know that this is not truly or fully decentralized but uh is a way to have a more decentralized network we have an amount i'm not we we decentralize until it hurts and we centralize it until it works and for now is the only way to make this happen and we also want to have the competitor of infrastructure as a service from the very beginning we try to incentivize people to run to run nodes because otherwise you don't have any reason to run a node or to have decentralized network and that is why having this model and having institutions to run these nodes is going to increase this decentralization right now a big chunk of the network is in the hands of aws digitalocean or other service providers and we are trying to figure out how we can solve that issue but we need the second piece of this puzzle this is how we are going to take these from the network and put in the hands of this system there are several projects working on this direction like little wicket has taken we are going to try to find a partnership with them we are approaching this problem from the base layer not from the top layer the smart contrast part but mainly we want to follow the same approach they are taking so mainly the idea is to take ethon on a down and create batches of 32 eth one cool thing is that we can also have several doubts it not needs to be a common doubt for everyone for example imagine that you have an entity that has a lot of it and want to split the yeast in several validators they can run this this dial itself and split it uh on the network so yeah well we take days we create batch batches of 32 e's and then is when designers need to generate this multi-party computation on chain to create the public key of every batch of 32-es when we have this pallet key then we can trigger the deposit of these deposits on the deposits deposit contract and at this point we are ready then to take them to take them [Music] to take the control of this validator and thank you to the management protocol we can assign this to the different dom nodes around the world yeah mainly with this we are allowing to take is in a decentralized manner um and to split the work in several entities and we have uh seen that this has a lot of advantages we have liquidity sticking but also liquid hardware it's important to mention that we are not going to need collateral to implement these solutions the only one that need to put collateral maybe is going to be the the consortium because at the end the workers that are the dominant that the that are making the querying work they are not allowed to make any problem to the validator process increase the decentralization levels um we don't have a withdrawal private key at any moment since the withdrawal is going to be the smart contract directly the dao we don't have a validator private key at any point since we are going to use this dkg ceremony to create the validators and we also integrate individuals to maintain the network because we expect that they are going to take a fee for making this work and another coolest cool thing is that we are going to be more geographically distributed hardware you know one of the key points of all this technology is to be resilient to a censorship and making this happen is a key point to make the network more strong for everyone this is our expectation uh the the robot for the coming months right now we are working on um on the slashing protection and also the communication with the clients and and yeah we expect to start to having uh first turn it very soon um but for that we need you uh we we have a very talented developer helping us like uh prada or nando and also jesus but we need more help working on this uh this is a important piece of infrastructure and we need more people helping us to to build it so if you are interested on this uh you can send me a private message or it's poll that is around here over there and we will happy to introduce you our technology and and make you part of all of this so well this is our current plans um thank you for for being here and if you are interested on know more about this you can reach us in discord or send us a different message so thank you so much