Alon Muroch 🛸 Improving Censorship Resistance Using DVT

[Music] hey everyone you can hear me well right with the audio perfect uh my name is Alan I'm one of the co-founders at ssvs uh core team uh SSV is the first implementor for uh DVT which is distributed validators technology uh I started out life uh actually in the etherum foundation is a research project the idea was how do you take an ethereum validator and being able to distribute it into shares and those shares operated by independent operators node operators uh with the idea in mind that if you can do that on a large scale you can actually decentralize a lot of the control for uh for each and every validator um and that's kind of where sensorship resistance or at least what DVT can contribute to that uh comes into play um I mean the state of staking on ethereum is is uh pretty well known for everyone it's uh it's pretty challenging there's a lot of centralization ation happening uh specifically because it's hard to run your own nodes a lot of people don't have necessarily 32 e and so what happened was uh a lot of that stake is being put into the hands of uh staking services and protocols and those staking protocols uh became very large uh and also the the actual uh companies and entities running that infrastructure became very very large um that causes a lot of issues I mean uh today you probably have around five or six entities unique entities that control about 2third of the stake on ethereum many of them in the same regulatory uh geographical location which is in of itself um not doing great for sensorship resistance um it creates a lot of challenges for ethereum it creates a lot of risk both for the users obviously whoever is taking a large uh staking service um is it's a big Target so they might be more uh impacted by hacking or even accidentals uh but also because the way slashing on ethereum Works uh if a very long large staking provider get slashed um the impact or the penalty becomes greater than just a single solo Staker getting slashed um and there's also systemic risk on ethereum itself uh amongst other things censorship uh or the idea that you can decide quote unquote which transactions or which actions you can do on the blockchain Just Because of You controlling a mass amount of validators um there's a bunch of different reasons why censorship can happen I want to focus on three of them uh obviously there's more um but there's always that malicious attack that creates censorship and we'll kind of go through what that means there's business decisions that call cause censorship and there's regulatory reasons that cause censorship all three of them and amongst other obviously um reasons why censorship can occur um are are dangerous because ultimately we want um an ethereum that processes everything is not um swing that way or the other um so a malicious attack that can cause censorship obviously starts with some malicious actor taking control of of of of a of an entity or a service provider that runs a lot of Stak so it can be through hacking it can be through internal uh just an internal work somebody did or if it could be even an accidental uh an accident somebody did within the same entity or same company um the censorship there might not be it could be both specific or very broad again it's it's it's mostly to cause Havoc uh and it's for the most part temporary uh hopefully temporary and so that's kind of the the malicious part of of of um of of some malicious entity third party or even internal trying to do harm uh and and part of that harm can be SL can be censorship um obviously that regulatory censorship or regulation based censorship is is something I think that our industry in specific has a lot of challenges with uh it can be forced or or self-imposed um it can be it's for the most part it's very specific so it's for very specific type of transaction it's not broad um uh and it's evolving it's it's something that starts out in one place and in a year two years three years will be a whole different set of of censorship uh called them filters or conditions and it's for the most part long lasting so it's it's not something temporary it's something that when an entity being forced or decides to censor because of Regulation it's probably a decision that will follow for through through years um and the last one is really uh the business case it has a lot in common with regulation but it's mostly driven by uh for the most part for Rewards or or just dollars um and it's for the most part self-imposed it's very specific uh it's still evolving and it's long lasting uh so for example you can decide as an entity I'm not including some transactions because I can't be bothered uh to spend gas on them or or or or just processing time or whatever whatever the reason is but all three of them ultimately lead to the same um non-desired kind of outcome which is not all transactions are born equal uh and that's a problem and I think there there's a bunch of different solutions I want to or or at least ways to mitigate that I want to focus on two um so one is it's something that I think for the most part we saw as a as a community being pushed around or not pushed around but being kind of promoted which is just increase the number of uh unique entities or people that run validators and actually run uh validators that process and propose blocks and so on and so forth um I I think it's a it's a great thing to strive for personally I think it's very limiting it's not very scalable um obviously that's something we want to see but uh I think we we should be looking for more practical type of solutions and then and then and the other one I want to talk about is DVT distributed validators technology um that's something I've been working on for the past almost three years and I think it has a lot in terms of value it can add or or mitigate some of the sensorship risks um and so what is really DVT so DVT is the ability to take an ethereum validator and split it up into shares and those shares operated by completely independent operators node operators um those are kind of connected through one another in a group we call them clusters uh clusters can vary in size so um uh it can be four seven 10 or or any other number of node operators within that cluster and DVT really combines two pretty proven Technologies so one is threshold signatures with BLS it's super super easy to do and the other one is just consensus protocol a thin consensus protocol which lets those uh operators node operators decide on what they want to sign and then use threshold signatures to actually sign it and and the consensus protocol really gives us liveness and and fault tolerance in general um just being very very robust against uh both offline nodes and malicious nodes and so on and so forth and so that's really what DVT stands for and what it what it really does um and if you combine additional kind of um layer to it through the key generation process if you use dkg for example then you can even take it a step further and say okay nobody holds the actual key and so there's less surface of attack um and and it opens up some kind of new properties specifically around regulation um in terms of responsibility and who really owns the key and the validator through through that for that matter um and so DBT with dkg I believe can bring us uh to something very interesting I I tried to kind of you know put it into some graphical way of of kind of showing it around but um this concept of censorship resistent cluster um is something I think uh is worth talking about the idea that you can take a cluster configure it in a way that uh really reduces the ability to censor uh through that cluster if you do it on a large enough scale then you have a big you can have a big chunk of the validator set um really 99.9% U sensorship resistent and so what is this uh sensorship resistant cluster so the idea is that you can mix and match different types of operators node operators um into uh something which gives a lot of protection so in this case what I did was I took four institutional well-known institutional staking uh Services um regulated some of them in the US um currently hold a lot of stake and their value is really that they're probably the most robust and scalable uh out there they can handle the the scale they can handle handle being attacked they can handle they've been around the business for a long time right that's that's the bottom line and so you have four of them and the cluster side is seven and so the three additional ones are just solo stakers they can be smaller institutions they can be actual solo stakers whatever whatever the setup you want the idea there is that you have seven different unique entities four of them could be potentially regulated and censor in their own way but because they're attached to three others which are private and completely uh different type of of of an operator those three only need just one more of those institutional stakers institutional um staking services to sign something with them in order to broadcast it to the beon so imagine that for example there's coming up some reg new regulatory framework in the US saying uh certain types of transactions will be censored well the solo stakers only need one more entity through the four others which are institutional to agree to sign with them a block and that's something I think pretty plausible in the real world not all of the staking uh institutions are in the US not all of them treat uh risk at the same way and so on and so forth and the and the flip side of that is that you still have the robustness of four institutional stakings staking powerhouses and so if one of those or maybe even two of those solo stakers fail because they're not as robust maybe they are even amateurs the others can recuperate because again DBT holds within it cons a consensus protocol which gives you full tolerance and so I believe that this type of combination where you mix and match different types of staking uh uh uh node operators and staking Services you can really make a long stride forward in at least on the surface and at least conceptually make sure that it's not as easy as today to censor because today really you know if again we're going back to the first slide and talking about around five entities controlling the majority of the stake in ethereum and most of them in the US it's pretty easy to censor uh the only thing that the regul regulator needs to do is just send them a letter uh I think none of them will say no to sensor a specific transaction uh although the implications are really big but imagine that the even coinbase like the most heavily regulated company out there imagine them in one of those clusters and imagine the validator key was created in a dkg coinbase can say are right that they don't control the validator they are part of a committee and they can simply outstand of signing that specific block where the specific transaction might be uh contagious for for them specifically but you still have six other entities who can sign it and the threshold is four out of seven and so you have that bandwidth to say okay maybe this one I'm sitting out as coinbase but the others can still continue in signing it and nothing happens on the surface and so my personal belief is that this type of approach is very very plausible uh and it's uh definitely something that can be happening today I mean SSV has been around for a few years on test Nets we launched on Main this is something that can happen even today on Maine no change to the uh ethereum Network at all uh signatures are completely transparent for the Bitcoin chain that for that matter um and so I believe this is a good approach take a lot of the risk out uh without major major engineering and changes to the network itself um like a few words on SSV SSV is is a contract based DVT Network so the idea is that you have a bunch of interconnected node operators and they're all synced through a smart contract that smart contract decides basically or or rather its users decide which validator is run by which cluster and again cluster is just a set of operators that run a specific validator um and it's um it's a it's a Global Network we're running on multiple data centers and continents and so on and because it's a it's a contract based Network interacting with it super simple just a simple web app and you can start running your uh validator I truly think that having the discussion passion with to integrate more uh unique Solo solo stakers or unique entities that run operators through something like DVT is a very simple way to take a lot of the risk out of censorship and I think the risk for censorship is primarily due to those three factors business regulation or a malicious activity and I think DVT handles uh and can help in a lot of those cases um thank [Applause] you again