Economic Mechanism for Rollup as a Service

Transcript

foreign here from earlier and for this session right I will share with everyone about the economic mechanism for roll up at the surveys actually it's a very new topic I believe a lot of people um like we don't probably know like what's the role of a service so I will start with sort of why we need a low level service and what kind of roll-up service can provide for us and after that we go into the economics step by step so as we all know right right now we are sort of in a paradigm shift from

this modular design to this um sorry monolithic design to the modular design previously like uh six or seven years ago we know that for ethereum and bunch of these l1s they are sort of like a monolithic chains uh basically including execution settlement consensus and also data availability layers together but right now as we can see right we sort of like split this entire test stack into different layers we can say there's an execution layer as different relapse and we also have the settlement layer as ethereum and sort of some other air ones and meanwhile since ethereum is turning

into this low lab Centric scaling approach granular we can see ethereum is becoming more and more like a data availability layer and all the execute fusions will be offloaded to different Joe labs and right now we have so much I mean better landscape for all different execution layers as we can see right we have Optimus execution layer like optimism every Trump and the various like the other Solutions probably not just evm compatible I mean well we also have different ZK Solutions like decreasing polygons evm which evm compatible but at the same time we also

have some non-evm compatible like style wear and at the same time right apart from the general purpose l2s we also have different like application specific ones like what we are building and also a different like uh research Labs a building like these execution layers for different Industries or sectors like game sector empty sector and the social applications and since we have so many like sort of new execution layers and for developers when they want to spin off a different channel app or l2s for the applications to cater for millions of users right sometimes they're quite confused like which L2 or which test tag we should

choose and we can see a lot of sdks but how to incorporate or integrate these sdks into our existing test stack and how to like sort of uh deliver these l2s and attach to different ones it's sort of like uh like sort of a little bit headache for the Developers and that's why we need this rollover service providers and this kind of companies and projects for Journal best service right we make it you can treat it a little bit like a software service for these companies like we incorporate all these sdks and also different test tags into

these kind of solutions so in that case for developers if they want to launch your Labs or l2s for ethereum you can just leverage some dashboard or existing tools and quickly spin off these l2s and posting data or deploying bridges on ethereum and as you can see right I even for Russ or developers service providers right now we also have a quite extensive like uh um like landscape of all the different solutions we have sdks we have a different sequencers a site provisioning and at the same time we also provide writing these different node no code

deployment as I mentioned right so right now we are already at the stage that compared to six or seven years ago we have to build from Strauss code or using some like sdks but right now we have a bunch of this dashboard platforms you can just choose the parameters you want for example Define the names for your job and then within a few minutes you can launch a roll up or L2 for you so and then as we can see right uh did a bunch of benefits we can achieve where your life service especially from different your lab service providers so why not is is really like flexible and low cost and fast deployment as I

mentioned right right now for most of the rollers providers basically we can quickly spin off the roll up depending on your requirements within a few minutes beyond that as I mentioned right for a lot of these developers from web 2 industry or gaming sectors probably don't understand like how to relaunch a blockchain and I'm also they know how to develop insolutely so in that case they need these kind of no code framework that's something we are providing and also a different rollover service providers providing um so as we can see right this becoming a trend that for a lot of developers you

can just call the API or their dashboard quickly spin off the L2 you don't need to worry about like sort of read the source code or like debug this sdks right now it's really like AWS in crypto industry you just quickly spin off these Rolex and l2s for your applications in that case right how we really monitorize this kind of solutions or Services as I just mentioned right we need this kind of rollover service to create a spin of the road labs and for our applications but how like these providers like us or the other companies make money so one straightforward idea a lot of us

meeting boys why not directly apply this SAS model into it for example uh when developers spin off the roll up right let's charge some fees directly either similar to GitHub right GitHub or AWS for GitHub it's sort of like subscription model right it's perceived if you have more Developers for your organization we charge higher fees for this deployment but at the same time um you know we all know that right for your lab typical it lasts for a long time and weekend share fees either month by month or like sort of like like year by year so in that case it's quite similar to the existing SAS model you

can connect to the strap and some other payment services and just charge user fees as a normal software um and in this model right the good thing now is quite straightforward we have a lot of mature solutions for this SAS model regarding payment side right we have a strap and a lot of different Payment Systems even crypto payment system from coinbase and some other providers and in that case you just need to input your credit card and sometimes you can pay in crypto and based on that as I just mentioned right they also tiered pricing probably you can have a starter tier developer tier and a corporate here for different tiers like

we can provide different Services started here we can provide some like total you can only have one seat in your organization at the same time you can only launch one Euro Lab at the same time but for developer tier or corporate tier you can provide much better like services like you can launch Five Road apps at the same time or cooperate here it depends on your demands right we can launch as many as you want um meanwhile they also have repricing and uh and for a lot of projects right at the beginning we can charge no fees like apply this a freemium way like probably for the first one or few months it's free for you if you really like to

use it and later and we can extend sort of the subscription by charging some fees and beyond that we can also provide credits and to attract the users or developers and later on we can charge based on the usage and apparently we can see that if we directly apply the SAS model is class three four right but you know right we are in crypto and your lab is quite a new area like we enter into so there are some complexities if you want to apply this SAS model so let's look at like what are the steps and what's the cost and what's the revenue we can have from these SAS model from

the zolab space and then we will see like how much like sort of Revenue we can get from it so this is quite straightforward your lab workflow overview right we have sequencers basically client transactions and aggregate transactions from from different users Aphrodite typical sequences are executor execute the transactions and composed blocks and basically later regularly there will be either sequencer or some meter submit the data to the L1 and typically they are also from proof of validity proof on the L1 if there is something wrong happening around the lo2 we have different ways on the L1 where the smart

contract to verify these proofs and to challenge these sequencers this is a typical way right but typically how we charge fees from the users for these L2 systems as we are using like probably optimism or decreasing a polygon ZK evm right we all know that when we pay fees it varies a lot depends depending on the time within the transaction so why the fees is not constant because it's not just like sort of the L2 it also includes some of the fees we need to pay for the data when the sequencer pulls the data on L1

so most of us specifically as you can see right L2 fees typically means the the like normal gas fee right like like what we can see from the recent launch linear the gas price probably just 8GB um it since since it can handle a lot of transactions typically the gas price is quite constant and but at the same time it will frequently like post the data to the R1 then like for uh during this uh period of time right it really depends on how many transactions for this period of time if we have a lot of transaction right we can sort of um like average these costs for every user but if we only have very few

transactions like 10 transaction then everyone needs to cover higher fees for the data post posted to the L1 so that's why as we can see right the fees we pay to the L2 is not constant even the gas price is the same so then as we we can see right so there are a lot of places for users to pay the fees to the auto but I at the same time how to ex actually is making money this is a chart from optimism and also happy Trump to record their revenue and also the the cost the costs are mainly about the poster the data to the L1 for most of the Autos we we sort of like charge

extra fees um we overestimate the fees and the charger user because we we may concern that if we don't charge enough fees later there's a sort of gas wall or gas surge on the L1 we don't have enough guys to pay the data we post on the L1 like ethereum so that's why if you look at the data right for optimism Arbitron actually for the past few months they already made a few million dollars basically the fees the declarative from the users and uh and then deducted by the sort of fees that they paid to ethereum for storing the data and then let's look at some like sort of like like some formula or some

um some these kind of uh uh like uh results we we compose from the discussion just now right so technically we charge the user fees based on the L1 data publication and also the L2 operator fees plus L2 congestion fees assume there are so many transactions like every second we need to precise a thousand of transactions then the operator cause is quite straightforward right the um the running the machine life is a machine from AWS Google cloud or barement tool machine plus the L1 data publication cost and at the same time right operation also has the revenue uh here I count the user fees plus the Mev

Mev is another topic probably um like once we have time we can discuss more about it and then the perfect is basically the revenue minus the cost and and basic as you can see right it's it's just like uh aggregation of the user's transactions and uh minus the L2 operation fees more specifically right if we consider the normal General general purpose roll up as one case right then as a rust provider if we provide the similar services to run a general purpose L2 then the cost and also the revenue and purpose we got should be quite similar so that's the

case as you can see right we have different like sort of rows consisting of fees and costs and in the end for the operator is basically we can collect the fees from the users and um and also some service costs service costs from the um the customers for example game company or nft companies and it just manage this cost and and also the the basically the data we need to post on the L1 but there's another case since it's for general purpose L2 right but for a lot of the l2s or applications being of digital lab they are quite application specific a lot of customers

talk to us like game like big web 2 game studios what they want the one zero fees zero gas fees they want to provide the user seamless experience just like like the web tool gaming so in that case right we probably cannot charge the fees from the users at the same time there's no such like Mev fees because they want first come first as well so in the end right for normal rust provider for some application specific use cases like games empties and probably social applications um the the overall per feet will be much lower than providing a general purpose roll up L2 and I as I just mentioned right

um beyond the transaction fees we can collect from the users we also charge extra service fees as I mentioned at the beginning right uh we can apply this SAS model on this rollover service so what kind of the SAS model we can have like this is subscription based or we can also charge commissions from the Ft sale or some tokens that's quite unique in your lab space or in crypto space compared to the traditional size and beyond that the other ways are quite standard right either it's a flight pricing or sometimes usage based like pay as you go similar to AWS or it's per user price in GitHub like

um like for each organization more members and we charge you more and then in the end right so um based on the formulas I mentioned I uh a top right so it's still like we can charge some user fees if a general purpose L2 if is the application specific and there's it doesn't charge any gas so that part is zero and the service fees we always charge unless we have some free qualities for the customers and beyond that there are always the cost running out to operators like for Optimus mainly like we need to spin off like machines on cloud experimental places but at the same time right we also need to post the data on

the R1 so that's R1 data publication course the D the initial D but for this part right we can also save some money by playing some here like for example right now there are a bunch of these Solutions right like eigen d and Celestia or we can even compose a DA committee and to store the data like this software because of validium so in that case like we can we try to increase the profits I try to increase the revenue by U and S but at the same time we try to reduce the O and the D that that's also the way like if we really want to apply this SAS model

into this rice space that's sort of the formula we get and uh of course right I I just want to share a little bit more about uh that sort of the cost based on our experience so for the last course as I just mentioned right the major cause is really about the operation for the L2 operators and typical we call sequencer verifiers and meanwhile they're also the da cost and a lot of people ask us right and so right now we already have the op stack a ZK stack and the various execution layers so what's your opinion on the course on this so based on a lot of like um investigation and also comparison across different solutions typically if

we are applying this Optimus lab approach right the course is much lower because we don't need to run a super expensive like for example 500 000 dollar machines to do the ZK proof that's the part we can save a lot of money but of course there are also three dollars sometimes we need to experience the long challenge period but overall as we can see right optimistic Road lab in general the cause is relatively low but ZK Parts like we need to cover mostly on the approval course that's the major cost for Optimist in your lab basically we just need to consider more about like where we put the data like for example

it's on ethereum or sometimes on the other da layers the state's goal for the Ross right as I just show you so basically we have these different uh pricing plans like uh flight pricing till the pricing and also different models but it typically highly associated with our use cases like I have Team mint gaming and sometimes it's um it's sort of like we make it uh free credits for a lot of Partners um and for the past few months and two years right uh either us and also some of our uh like like projects we did a lot of these applications via the rust

like empty camping we also support a bunch of these football matches the this ft ticketing and I mean well a lot of game uh plays like that Forest so these are the real world use cases and for most of them right it's quite straightforward just like typical SAS payment model and a very interesting one like you you can just scan the barcode this is something we did recently we provide the service and basically everyone can play it it's fully launching every transaction every move is a transaction on chain but at the same time we make it really latency list so basically you cannot experience any latency for this one so these are the

use cases like sort of we try to apply this SAS model of course like today I also want to show something quite different for the traditional SAS model right that's a way like we try to decentralize the entire landscape and try to make it as decentralized as possible so that's how we have this Beacon layer or we call this instrument it layer for your Labs you can treat it as like orchestration layer and when users want to spin off a roll up the directly talk to the beacon layer and send requests and begin layer will assign dedicated like sort of operators for this clients and they will spin off for these applications the cool

thing that as you can see right the first version we provide is real Rush dashboard basically you can spin off very very centralized way but later on we have this Beacon layer we make everything decentralized so people like all the audience here we can join as a solo sticker or solo sequencers into a network and later on once we get some like sort of Demands and requests so the product will come quickly spin off like the validators from you and to solve these applications and as I just show you right compared to different rice or SAS model now it's

quite decentralized protocol you send the requests and this Beacon layer will assign some of the commuting nodes and to serve these users and meanwhile it also provide like a decentralized sequencing decentral verification and meanwhile as we all know right if we rely on the security on the L1 there's a validity proof and also flow proof sometimes it takes it takes a little bit of time right but with this decentralized like Beacon layer we can quickly do the verification and then you can have this harder than soft finality uh beyond that compared to the traditional um Ross slash model right now since we

have the decentralized protocol and everyone can join as a node and then you can provide the service for different applications as the rollout knows right and then there's a bunch of ways we can make it right like a normal and a very uh decent economic flag wheel for this decentralized protocol because the ethereum community right we always prefer decentralization compared to centralized like sort of service so that's why we really make this brief move to really decentralize this Ras platform into this um uh like sort of a decentralized protocol um and uh yeah and then at the same time

right we are still having a lot of open problems for example it's no longer size right then for decentralized protocol how we get enough nodes and to serve different applications and I mean well how we capture the Mev and during all the different applications across different nodes and beyond that some application wants some customization right since it's quite decentralized how we make the customization possible and as like sort of uh um like Justin just mentioned right uh right now we have some restaking Services how we apply the risk thinking Services into this uh decentralized Rush model so we can really enjoy the

benefits from ethers so that's all the open questions and we hope we can together with all the audience and the ethereum community we can explore all the different new ways to really decentralize the Russ landscape and meanwhile to capture more values for ethereum thank you [Applause] thank you there you go