Revolutionizing Interop Towards Proofs By Bo Du From Polymer Labs

[Music] hey everyone uh my name is Bo I'm here to talk today about uh interoperability and the journey from uh messaging to proofs uh thanks for coming this afternoon to to listen to this talk I'm I so I want to frame the conversation here with a little bit of a story so I want to start with the story and I want to end with the story so imagine a mysterious portal appears we walk inside and on the other side we arrive in 241 BC China in front of us we can see the Fay Army forming and assembling and we watch them set out the battle the general Shin in this case is leading his Army against their neighboring countries the battle itself is fierce and chaotic the Fay Army ends up getting uh moved around all over the place and is losing the battle and they need a way to reorganize the shin General is faced with a choice either he commissions a few Horsemen to send a few messages out to notify the troops or he sets off a smoke signal to notify the troops on what to do we'll find out what the right answer is at the end of the presentation so a little bit about myself um my personal background is in database engineering I worked at a few different startups uh before founding polymer so the ethereum ecosystem today is expanding rapidly uh we're seeing tremendous growth in the number of number of rollups unfortunately for the lawn of rollups uh connectivity is major issue uh we estimate this problem to be only less than 10% of rollups today are densely connected and this problem is getting worse over time if we look at the historical timeline of blockchain uh starting from uh Bitcoin ethereum defi sparked a era of uh L1 expansion for a period of time uh although a few Roll-Ups launched sooner than 2022 uh in 2022 we had the launch of the op stack arbitrum Nitro and rollup Frameworks that ushered in an era of Rapid rollup expansion soon we'll see the advance of new Roll-Ups that are high performance like Mega eth rise and soon and so on uh additionally during this period of time uh you had a lot of different interoperability protocols starting with IBC Wormhole Axel layer zero hyperlane and now polymer launch uh to solve this problem many of these protocols have not been able to keep up with the expansion of the in the number of rollups I think hyperlane does the best job uh but still with an inefficient interoperability model so we can all agree that efficiently scalable interoperability is the key to ethereum success given this rapid expansion of Roll-Ups and the rollup road map itself well messaging works and is the uh interoperability mode that's taken by interoperability protocols today messaging roughly looks like this you have an app Gateway on the source you have an app Gateway on the destination uh an application message is routed through a Gateway usually you have like a gas Oracle component an offchain relayer and messages get pushed through the system messages in this model are explicitly sent from origin to destination uh this model is generally uh opinionated on communication pattern meaning it's generally Point too uh there's a lot of gas consumption both on the source and destination uh there's increased complexity latency and less flexibility overall uh so I want to pose a question of uh is interoperability is messaging required for interoperability with chains being verifiable uh we believe that sending a message is unnecessary with verifiable data instead of sending a message we can instead just generate approv proof a proof that something happened on chain uh whether it's a proof via an event whether it's a proof via the App State uh we should be able to prove things on chain polymer is working on a proof-based interoperability system that obviates the need for messaging proof proofs provide a simpler and more efficient solution the mechanics of it are similar to Native interoperability Solutions such as super chain native interop in unlike messaging proofs or Ed on demand and the application only needs to integrate with aover contract on the destination chain for verifying these proofs uh this in turn has minimal overhead when compared to messaging so if we scroll back we can see that all of the Intimidator middleware uh gets removed for just a fetch a proof fetch from the destination so what are some advantages of proof based interoperability proofs offer simpler user experience uh instead of sending a message the developer can fetch a proof and then use that proof anywhere on any chain it's a simpler integration the uh application Builder doesn't need to integrate an app Gateway contract uh it only needs to integrate aover contract on the destination and there a simple messaging format the application either just emits an event event or reads a portion of its state proofs are also easier to maintain at scale with messaging you have to manage the routes between chains and in the diagram on the screen we can see that as the number of chains grows uh the complexity of managing these routes uh grows as well we also have to deal with uh gas Oracle updates many protocols maintain an onchain gas Oracle that estimates gas for connected chains uh on chain and this would have the same complexity in terms of in terms of management proof scale a little bit more efficiently instead of scaling a bunch of routes across these different chains you can slap a single API in front that can talk to all these chains and request proofs event proofs or storage proofs and the topology of this setup is far simpler proof also give apps more control proofs allow apps to choose between broadcast or point-to-point messaging mechanics and gives apps more control over the guarantees made by the system itself the app can now determine how fast to go between chain to chain it doesn't have to wait for the messaging system to determine if the uh gas price or if a route is profitable for them or not an app can decide whether or not guaranteed delivery is is a desired property and the app is also allowed to bypass censorship meaning that the app can directly Force proofs through the ly native Bridge um uh proofs are also more efficient messaging is a sum of send receive and verification costs the estimated gas paid to the system is a sum of receive and verification with proofs the only cost or overhead is the verification it's self so what are some examples of proof-based interoperability we have storage proofs uh which is used in RP 7755 by Bas or arbit trim's cross chain broadcaster standard with storage proofs what we're doing is we're proving a key value pair in uh rollup on in in a rollup and then proving that rollup State on the layer one and the layer one state that's derived into the connected rollup on rollup B so if you look at the flow it's the key value pair to roll up a state to L1 state to rollup B State then rollup B can fetch a storage proof proving this entire uh chain of States all the way into its its own State while this works uh storage proofs can be quite slow as they go over the rollup native Bridges themselves for example in the diagram here we can see that uh to generate the storage proof at for the first leg it takes 2 seconds for rollup a to produce a block 7 days for it to settle and roughly 12 to 16 minutes for the L1 state to be derived into the roller B State overall this process is uh takes a week or longer additionally these proofs are expensive uh the size of this proof can be over 16 kilobytes and the and the gas cost or the execution gas cost can be over 2 million gas polymer takes a different approach uh we take the approach of using event proofs instead of storage proofs this means that the polymer rollup is responsible for ingesting the block headers of all connected rollups which from which we can prove events against a header itself in the receipts route in this model it takes two seconds for rollup a to generate a block header and another 2 seconds for the polymer rollup to verify a sequence or pre-confirmation from the connected rollup roughly about four 3 to 5 Seconds to generate a event proof over polymer uh these proofs are also much cheaper they're less than a kilobyte in size and cost less than 60 60,000 gas to verify in comparison to uh over 16 kilobytes in proof size and over 2 million gas to verify now what does polymer look like in production I pulled this example from an existing messaging protocol to compare what the costs look like uh live between two different rollups in one example that we looked at the latency of sending a message from chain a to chain B was roughly 56 seconds in an arbitrary messaging protocol in comparison to polymer the latency of fetching a proof and proving it on the connected chain is is only 5 Seconds this makes proof in this one example 11 times faster than messaging what about the costs when we looked at the dollar value cost of a message sent uh between rollups uh we can see that in the anecdotal example that we pulled a single message cost almost $3 this includes descend costs receive costs and verification costs which are bundled in the estimated gas paid to the protocol uh in comparison polymer proofs cost roughly 15 cents this results in a 95% s uh cost savings compared to messaging and when we compare it to a system that's uh communicating over rups that are using Alda uh the costs are even cheaper the proofs can cost a cent or less uh to verify and this results in over 99% savings uh we have many partners building with proofs on Main net today uh Eco open for super form Everclear colani and and and more low latency cross-chain proofs are only part of the problem a part of the solution to the problem of interoperability uh like they say Teamwork Makes a dream work and there's many great teams contributing to trying to solve this problem we have arbitrum optimism espresso polygon uh across and many others okay so now back to our friends in uh 240 BC China these are probably one ing why this story and how this story is connected to to the problem here uh the analogy that I want to make is proofs are similar to uh sending messers on Horseback I'm a general I want to communicate with different parts of my Army in real time uh I can choose to send a bunch of Messengers on Horseback to communicate to each of my units or I can set off a a colored smoke signal to indicate to the Army on how they should reorganize and what they do the recommended Choice here is to just send off a smoke signal uh I can communicate to the entire Army at once I don't need the overhead of a bunch of uh men on Horseback uh the latency is just faster the overhead is cheaper um and I'm able to do this at at scale across a very large battle a very large army so the so the Fay Army sends off a smoke signal to notify its troops to reorganize their army was able to regroup and the Fay Army was Victorious if you're interested in building with proofs or crossing Pro proofs if you want to build a cross chain apps whether it's intent-based application or just a multi-chain defi application please reach out to me uh we have a website that's listed here and you can find me on Twitter at zero EXO thank you