Privacy in Cosmos Live: Panel - ICS, Mesh, IBC & Privacy

Transcript

Okay cool so I think we've got a very broad uh panel title uh which is just disappeared but um so ICS mesh IBC and privacy and I think most of those would be enough for a panel on their own uh so we're going to try to maybe narrow the conversation a little bit um but basically I think broadly we want to discuss Shad security and how that's affecting the cosmos ecosystem and I think that's a topic which is clearly very important in and will continue to be for the next year for the ecosystem and maybe touch on to how that is evolving and how it will impact things like privacy as well

so at a high level you know the things in the title ICS mesh um Celestia potentially as well all have different benefits to builders who are coming to building Cosmos and so I think just to get us kicked off it'd be great if you could each introduce yourselves and then also maybe say which part of that overall conversation you're most interested in I'm Sam Hart I am the head of product and strategy at skip which is an Mev company and I previously worked at the ICF for about three years um I'm not gonna lie this bit of a scary uh combination of topics but um

uh I guess one thing that I can speak to is um I've done a lot of thinking about um uh what security means um and um in my mind basically that's uh you know there's several different facets really you're trying to create some kind of accountability Loop where there's there's like a form of verification and some some form of um uh punishment or or way of enforcement and with that framing I think you can you can maybe like reason a little bit more

clearly about the interaction between privacy and and security hi I'm Gabriel I'm a researcher at Celestia and teamlet of roll kit um I'm excited about shared Security in the sense that Roll-Ups that are built on top of Celestia can uh use Celestia as a shared security layer and it's interesting how by separating the stack into modular components um the things mentioned here like interchange security and mesh security

can interconnect and use components of the modular stack to increase the validation for example let's say we use mesh security and they want to have objective fault to validate is that there was a invalid State transition function um we at rolekid are developing prop rules for the cosmos SDK so mesh security could use those fraud proofs to prove to each other that hey this ventilator was lying and slashed them accordingly so there is a lot of interoperability here so it's very

exciting to see that space um I guess my like overview of is that privacy fits really well into like The Sovereign app chain model with its like own staking token the concern is that you know is that like what will liquidity look like for that staking token um and a lot of the like so there's like a like so like Nevada is launching soon from the enoma team um and like it's like a privacy pool for the entire Cosmos ecosystem

um but it doesn't really have any um robust let's say native value accrual mechanism um uh and so you know you could end up in sort of the situation where you know Nevada is securing a very large amount of value in the IBC and wider ecosystem for a privacy point of view um but the chain itself um has like very little Economic Security so you know it would be ideal to be able to like use privacy together with you know a shared security mechanism um to make it easier to deploy privacy related Primitives um but uh there's also like enormous challenges to that like you know a lot

of these things uh leverage um you know the actually having a validator set for like threshold encryption Etc um so like it makes it hard to like Port that into like a roll-up World easily uh you have uh the reality that uh like you're not running like cosmosm Etc on these chains um and like the staking modules are different so like interacting them with like ICS or Mass security is non-trivial um so it's like what is going to be the common denominator there to sort of secure privacy enabled apps and today is like low liquidity environment um is an interesting challenge okay yeah it was some really interesting um touch points there just before we

dive into any of those it would be great if someone wants to actually Define maybe for people who are less into the cosmos ecosystem what actually is uh ICS and what is mesh um I don't know if anyone okay yeah lucky go for it I can do it um so um ICS interchange security is a product that is deployed to the cosmos Hub right now primarily developed by informal systems and mesh security is a kind of variant or different take on a similar concept that is primarily being

advocated by the osmosis team and has a has a kind of collaborative group that's working on it um they're very similar they are both forms of staking so in the cosmos Hub case the atom token is is kind of reused or leveraged for the in order to stake on a secondary chain and in the mesh case there's kind of a bilateral or multilateral relationship between chains where both tokens are used to secure both chains um

and there are some slight differences in um you know the details under the hood in terms of how delegation is used and how the the kind of stake is split up how faults are accounted for and slashing works but the the general idea is that's a form of staking that is used to economically secure a distal chain and and really it's the the main distinction is this like uh this kind of parent to child versus multilateral relationship awesome so I think um maybe it'd be interesting to start

back on the the top exactly waste of um actually if you're if you want to launch for example a privacy preserving application whether that's a private stable coin or a mixer or something else um one of the issues you actually might face is that it's difficult to get enough security for your chain because you have low liquidity on uh maybe the staking asset or maybe validators don't even want to run purely that chain because it exposes them to regulatory risk or anything else so I think it's going to be particularly important for private privacy focused applications to look at these alternative um security models um for each of you what do you think

will be some of the most um sort of what would be the most interesting model for private applications and for example with ICS um you know would current validators on the Hub be willing to run um these privacy chains in all cases and how do you think that will play out starting yeah maybe exactly sure um I think all of these things represent big challenges and areas of uncertainty for the adoption of privacy and Cosmos is just that like is probably the most straightforward way of of looking at it um I think like one of the questions

would be is like what's you know I think one of the things that will be will be there is like um essentially like what API do people have to implement in their in their privacy enabled application in order to even potentially opt into one of these systems so like the the like consumer chain which is like sort of the child uh chain of the uh and the security side which is where a privacy preserving at though I'd say wanted to use the ad in validator set um would sort of live um it's probably a pretty simple API to implement on the whole um it's probably like much simpler than let's say a staking module to implement so maybe we'll see you know it's a lot

you know uh like that might be like an initial low-hanging fruit but then it runs into the political uncertainty problem which is like the political uncertainty problem is like will will will the cosmosup vote for these things will big institutional stickers like you know coinbance and binance vote against it for you know regulatory reasons um and I don't think there's a single way to know until someone tries interesting yeah um one of the ways you could circumvent that if you don't need for example threshold decryption for from the valve data set is to try to build your application as a rollup

um because Celestia decouples data availability consensus and execution you can focus on the application layer so now new privacy applications have to bootstrap their own valve data set and create it's an overhead that hinders people to develop new applications and iterate and maybe go through governance proposals and one of the other ways you could do it is focus on the application instead and build a roll-up that is focused on privacy enabled through avci and like IBC and delegate data availability and consensus to a data

availability layer like Celestia did you want to add anything or um so I I have had a couple of conversations about um so one of the other challenges with a privacy chain is um it is more difficult to fork in some instances and forking is one of the primary accountability mechanisms of a of a chain basically the threat of a fork kind of maintained some political equilibrium

so maybe I just add that um this is like a another dimension that needs to be considered and it is one of the uh is one of kind of the key security facets of a restaking system so um yeah but people just kind of put that into the into the air why is um is it hard to Fork privacy chains uh because uh State you you can't Fork out a single balance um so if there's a misbehaving validator unless you've pulled out their their

balance specifically uh there's no way to kind of remove them from the token set yeah essentially like edited balances like so like the notion of forking is like to a certain extent like there's this political threat that exists in transparent Cosmos chains which is like you know the the prop 16 Juno threat which is like uh if you if there's a if social consensus moves against you with a sufficient amount of vigor um there's a a like you might end up on like a a sort of uh non-minority Fork that doesn't or minority Fork that

doesn't uh um where and like the majority Fork won't have your balance but in a privacy preserving World um there's no possibility of doing like prop 16 right like you can't like balances are all private they're all hidden no one knows um no one would be able to like easily to edit the state to remove so on one hand you get like stronger property rights um from privacy which is like I think a good thing um but like it definitely changes the like social equilibrium cool so it seems like we're we're in some sense waiting to see with some of these models you know and see them tested we have seen Neutron now go live as the first um so I ICS uhcs

what have we learned enough from that have we learned anything from that and what would we need to see happen in order to have more clarity as to how these different models might work together some of the I mean it's it seems that like some of the concerns about at least in the short run about validators not running the neutron chain um uh you know that was like the biggest uncertainty with ICS was right would uh especially like sort of like the low the low like there were two categories of of validators who expressed a lot of where there was concern was one it's like the low engagement exchange validators were there were they going to run the neutron

chain um and then also you know um more regulated validators you know custodians Etc who expressed concern about being forced to you know run a new chain um that maybe wouldn't uh fit within their existing regulatory models we seem to have at least you know made it to Step One um in that World um and that seems to have gone okay um and I think like I think there's going to be a lot of uncertainty about this until for quite some time um and like a a privacy enabled chain that gets through governance I think we'll continue to uh we'll we'll continue to ask novel questions I think it's

maybe a good problem to have if we're asking that question like a privacy chain has been so successful that it it needs additional security to to ensure that you know it doesn't get taken down by a malicious entity um I would love to have that problem actually um yeah I think one of I guess like one of the other things that hasn't come up as much as like all of the like so like we act as if like mesh grow ups and uh and ICS are like like defined points whereas what the actual reality is we're going to see a lot more there's a lot of possibility

for recombinations so like the existence of a fraud proof or a validity proof um and it makes it a lot easier to for instance have like subsets of validator sets uh run a chain rather than having the entire validator set so you could end up like especially in a world where like you have like uh like sort of uh roll-up friendly privacy Allah um like what uh Aztec is building like that kind of a system could be used to do a system that was like a subset of the Hub validators run it right instead of all because it's like fraud provable state but you'd still get a validator set and be able to do threshold encryption um so like not also like you know all these things can be recombined and in

new and interesting ways they're not like discrete points that are like exclusive yeah and maybe it's worth at this point like teasing out I guess how I think about like a roll-up model versus um uh interchange security or mesh security model like basically Roll-Ups um we put a lot of emphasis on the verification aspect of security um you know DK roll ups you can prove that the state transition happened and mesh security interchange security put a lot of emphasis on enforcement so if something happens we can punish misbehavior so there's actually a high

degree of complementarity between these things a the fraud an optimistic Roll Up is kind of a hybrid between these because there's a an economic bounty in order to ensure verification but um yeah but it just kind of speaks to the the expansiveness of this design space like you can combine these interesting ways and and we can actually push quite far on the verification side of um privacy chains by having like clients that verify the entire state for instance

so so there's there's different ways to increase your security level um as a sovereign chain how do sort of data availability concerns tie into that problem if you if you want to be doing using live plans for for example proving slashing conditions between producer and consumer chains or any other sort of variant [Music] so um with Celestia you can delegate the data availability and consensus part where a rollup can post their transaction data to Celestia and the celestial Elders agree that those bytes exist

and if this Roll Up is also run by an ICS chain where the validators are producing the blocks for the Roll-Ups then the data availability can be still on Celestia and those validators would execute those transactions and produce for example the block headers so you could have a combination of Economic Security that is for approvable or that you can have availability proof on that and like use both scalability mechanisms here cool so something to touch on uh towards the end how

um will all of these and how do how do you imagine Mev uh sort of evolving in the ecosystem with these different types of models maybe it's a little bit too early to really be seeing how it will change but do you have any sense of what that could look like um maybe exactly first well I mean I think one of the most robust mechanisms for just eliminating Meb is having privacy right if uh if the block proposer doesn't know what the transaction is going to do and can't simulate the future State transition um they can't uh easily they cannot they can't really extract Mev from the system so I think uh or like some types of Mev will can be can be mitigated and go away

um there's definitely you know there's still like being able to like adjust your like liquidity positions uh at top of block and stuff like that which might be which would be valuable um and I think the Celestia and the world kit teams have done like an enormous amount of thinking about different models and where Meb uh accrues there um yeah yeah um skip has been thinking about this a bit it's a little bit early but um as Accu was saying if you are hiding um order flow then um there is less opportunity for economic leakage

in those transactions on the other hand um typically the decision to to hide order flow has you know some trade-offs here you're you're going to be batching or introducing noise into the system so just to set up trade-offs there um there's another dimension to this that we've we've actually put a bunch of time into which is um uh in the you very likely want um privacy at specific points in the supply chain um really for accountability reasons so I for instance it would be helpful if

validators are unaware of the incoming transactions at the moment that they receive a block um so they have no um incentive to withhold that block and try to extract value in that way so that can help with the the kind of system robustness um yeah there's a lot of directions to go with this but I think it's very early days and privacy will likely play a role at various points in the you know in the flow from transaction from intent creation to uh to execution

yeah and one of the nice things about building this ecosystem in like a plug and play way where one entity can focus on one thing and people can delegate that as a service is also if you start looking into uh for example shared sequencers where a roll-up will be able to delegate transaction ordering to a shared sequencer and now how does the the sequencer or those transactions and do they maybe have some kind of Mev protection or acceleration or like depending on who does what in what kind of way you can increase or decrease Mev and

redistributed which is very exciting because the the modular design space explodes the moment you have very clear roles and those roles can be explored in their own way awesome thank you all very much for the panelists I think it's clear that as you in a way it's almost a challenge because the design space is so huge to really explore all the different routes this can be taken but the best thing we can hope for is that one or multiple privacy focused projects actually launch and and use these mechanisms and start to really give us Clarity on how they should and can be used um but yeah thank you very all very much

for your discussion [Applause] foreign