rom November 18th to 20th, FilSwan Founder Charles Cao was invited as the judge for the Hack FEVM 2022. At the Hackathon Kickoff, Charles gave a talk titled “Unlocking DataDAOs with FEVM: Lagrange DAO”.

Hack FEVM (the Ethereum-compatible layer of the Filecoin Virtual Machine) offered builders from around the world a glimpse of a nascent project at the bleeding edge of the Filecoin ecosystem — the true frontier of web3. Unlike previous events, Hack FEVM gave developers a unique opportunity to build on the new, Ethereum-compatible layer of the Filecoin Virtual Machine.

Unlocking DataDAOs with FEVM: Lagrange DAO

Jacob Willemsma:

Charles is working on another Data DAO site, which is the Lagrange DAO. Charles, welcome to the stage, and feel free to start your presentation when you have a second.

Charles Cao:

Hi, everyone. I’m Charles, Founder of FilSwan. We have been working on the Filecoin storage and development ecosystem for quite a while. Recently, we have found new opportunities and started a new journey to work in terms of the Lagrange DAO for the data sharing and analytic space for decentralized scientists.

Why ‘Lagrange DAO’?

There is a lot of interesting scientific data in the world that needs better services. And there are lots of good public datasets as well. People want worldwide access to these datasets and a lower cost of computing power. They want to have contribution and social impact, because when they do the data uploading, they want people to know that they could do something good for society. Also, there are a lot of things on the data, but they are not really on-chain. Sometimes we are confronted with link loss, space loss, and work loss, which is harmful to those precious data. And at present, basically the data version is all based on the centralized system. And sharing the data is based on Google Cloud or AWS Cloud.

So, there is really a huge market. From the graph, you can see that the storage of scientific datasets will amount to more than 250PB after a few years. This means it will cover 10% or even 50% of the Filecoin storage, which really means a lot. Moreover, there are different scientific data hosted in some centralized storage places and are looking for decentralized storage spaces.

We really need to build a place for decentralized science without borders. There are a couple of different places where people need storage and decent incentives. Enterprises have their data stored in hard drives. Storage providers are working hard on better-improving storage technology. Developers are hammering out the retrievals. And those token holders, with Filecoin tokens in hand, want to encourage the ecosystem but do not know how to do that. Therefore, we need a better place to add different people in the same cycle so they can contribute to the bigger system.

Data DAO in FVM

And with FVM, we found a better opportunity to get everybody together. We can convert the large amount of data we have to data schema and store it on IPFS and on the Filecoin Network. Previously without FVM, the storage lasted for one and a half years, which is an 18-month storage. But now, it is possible to make an extendable long-term contract to satisfy long-term storage needs. For example, scientific data in universities may usually need more than 5 years or 10 years of storage. Therefore, compared to what happened in the past few years, it is now a good opportunity for universities to store these data on Filecoin.

On the workspace side, we can create a place for people to share their data and models. And in terms of payment, ZK proof can be aggregated on FVM to reduce the gas cost and increase transaction volume. Cross-chain payment with FVM integration gives you more possibilities to make a payment not just in Filecoin, other blockchain users on the Polygon chain, Binance chain, and more, can also make payments by using the cross-chain DeFi bridges to enable the storage on the Filecoin Network. FilSwn has already done some work on facilitating that cross-chain payment, but now with FVM, it becomes easier.

On top of that, we can also work together on modeling and computing. There is a lot of computing power existing in the Filecoin network for sealing storage, but it is not 100% dedicated as sealing. After the storage is onboarded, some of the computing power can be shared for cross-chain computing. And if we can we create a workspace, put the dataset and model together and add……

To realize revenue sharing among data providers, storage providers, data consumers and computing consumers, a token economic model needs to be formulated. We need to make efforts on data onboarding and allocate dedicated funds for those who are willing to provide storage and who are willing to use the data.

Data DAO Tech Stack

This is the technical part of the stack. In order to make a stronger system. We need to use the smart contract and core services to make the payment while we upload the data to the Filecoin Network. You need different people to swap the tokens so the external ecosystem, they can use smart contracts based on DeFi cross-chain to use the same storage services. In the end, you need to give people the capability to mint their data assets as NFTs. It will be great if it’s a one-click experience, instead of writing a whole bunch of code. And in the end, all the resources in the Filecoin Network can be contributed to different systems. The critical part is data governance, computing, and validation. So, we need a data Oracle to help prove the data on-chain, data availability, and data computing capability. This part is done by a DAO organization. You need a multi-sig wallet to sign the proof on-chain. We have done some work on top of that with the help of Chainlink External Adapter and Filecoin native FEVM.

Data and Payment Flow

This is the data and payment flow. We found that there are new business opportunities in the new DAO model. For the existing Fil+ system, there are a couple of things that need to be mentioned, like the data providers. The dataset owners and the model creators are data providers. And in the FIL+ project, the deals proposed become Fil+ deals, which value at 10 times the computing power. The storage providers are willing to store the data for free, and even, they are willing to share some of the revenue with the data providers.

This generates some confusion: why storage providers sometimes are willing to pay for storing the data. That seems not common in traditional business. This is because when storage providers start to use the data to earn extra FIL, they become data consumers. From a business point of view, data consumers need to pay for everything, including storage, computing, bandwidth, etc. However, when a storage provider tries to get the Fil+ deal for storage, he is downloading the data from the data provider. And that’s why he needs to share the revenue with the data provider. It is for data downloading. Data is the value. That is how it works for today. Nevertheless, there’s no clear way and a clear market about how we can realize this process automatically and through the smart contract.

Data Schema/Model Storage

This is the data schema and the model storage. When we have a big dataset, for example, more than 400 TB, it will be trunked to 32GB data pieces. The index of the data should be stored in a safe place as the data schema and this scheme can be stored on IPFS because people need the index to easily fetch the data relationship. At the end of the day, those data, including the data schema, needed to be backed up to Filecoin and to 5 nodes, or even more, to provide enough redundancy and availability. The data space is also stored on Filecoin, because it’s huge in size. And in the end, you can trade the dataset, mint it as NFT in the data market to show that you are the actual data owner of it and that you want to provide opportunities for others, and make the funds together.

This is one example from the community that used the same methodology to mint the Filecoin snapshots as NFT datasets. Back up the Filecoin snapshots on a schedule, store them on the Filecoin Network, create a CSV format and JASON format dataset schema, and then mint the schema as an NFT with the smart contract and put it on OpenSea so that everybody can fetch the dataset schema from the smart contract. And This is important. So, with the key pointer in the smart contract with the IPFS hash, you get the data safely preserved in the network for a long term.

And in order to make computing ultimately easy. We need to create a space that can put the data providers and model providers together with the computing power. And in the end, the community will become a computing result backed up by the Filecoin Network.

Swan Web3 Suite

We have done a lot of work in the past few years, and we have a website suite to worldwide search all the available storage providers and give them some credits. So, we know who is good for the community, to store, and who is reliable. Our cross-chain storage solution ‘Multichain Storage’ has already been deployed on the Polygon Network, and we are planning to migrate to FEVM. Multichain Storage is solid and compatible. It can automatically back up the storage to 5 nodes on the Filecoin ecosystem. And we also have edge computing nodes that can do the computing based on the IPFS content hash.

Now, we call for participants to join the ecosystem during the revenue sharing. We have an estimated revenue of about 5 to 10% that we have shared with all the dataset uploaders.

We have been contributing to the system for a long time since 2020. We participated in the Slingshot V1, Slingshot V2, and Fil-Austin and held the Data DAO Hackathon, from which we have attracted a lot of ecosystem projects working on the data and relative building.

‘Lagrange DAO’ Overview

Now I’m going to show some work that we have done already.

This is the system we previously used as the V1 draft version. We can connect it with the Filecoin Wallaby system and can log in with our MetaMask wallet. You can view your wallet address and your FIL balance. You can create your datasets according to your needs.

If you want to create a new one you can just add the dataset. You have your wallet as your Owner. You can give the name of the dataset, for example, eth.globe.

And you can choose the data schema. For example, I have a data scheme here already. You can put the data CID here so in the future you can fetch it. And here you can create a new data CID.

We’re still working on the model and space and computing. We managed to get the computing work here from TensorFlow, running everything and doing all the computing work here, and getting the result from all the systems. So, the next step of the work is to make every piece work together, including computing hash, machine learning, and IPFS integration, and create a workspace and a token-sharing model for everybody.

FilSwan Product ‘Multichain Storage’

With this Multichain Storage, we’re working on providing people with 30GB of free storage. So, with those storage, it can be connected with the Data DAO. You can use this as an AWS-compatible filing storage. And the data used to be stored on the Data DAO will be backed up to IPFS and then backed up on 5 nodes. And those storage providers will share their revenue with you. And as a data provider, you’ll get income, because your data values. When data models start and computing powers start, you will gain extra value from your machine learning models.

That’s pretty much about what we are now. And we’re working toward the next milestone in bringing computing power online. Please follow us on Twitter. And if you have any questions, you’re welcome to talk to me on Telegram. Thank you.

Jacob Willemsma

Awesome! Thank you, Charles, for the presentation and for those demos on what’s been built. The underlying infrastructure makes it easy for end users to store files or then eventually do compute and more. It’s a really valuable effort that’s being put into this. Thank you for that. And I’m curious on your end if you were going to build this weekend at this Hackathon, what would you be thinking about and what are some of the projects that come to mind that you’d like to see built this weekend?

Charles Cao

I’m really interested in the projects here, like making use of the data. We saw a decentralized routing project was built in the last Data DAO Hackathon. The CDN network is building on top of the data. They tried to use the revenue-sharing from the data storage and redistributed to other services, become a cycle. I’m also curious about how people will design the governance part, like voting, sharing opinions, making proposals, and submitting. Also, we’d really like to see people use IPFS with Filecoin and provide value for access. Like Mirror, we know that Mirror is already using IPFS and Filecoin storage. But how can we make them use more and eventually pay more in the future, that is a very interesting topic.

Jacob Willemsma

Cool. That totally makes sense. And someone asked is there really a connection between IPFS and NFTs. I guess it might be worth dipping in a little bit closer into how you can use IPFS to build that storage layer behind NFTs.

Charles Cao

Yes. Actually, NFTs stored on IPFS are valuable. But the only issue is that a lot of NFTs have IPFS content hash, but they don’t really tie together. And where are they stored, right? That’s just on IPFS. How about the IPFS content? We know that IPFS is not incentivized to store for a long time yet. So, have IPFS storage backed up on Filecoin and build the relationship and bond them together, then users feel safe, because storage is on both IPFS and Filecoin, which is permanent storage. That is how it works.

Jacob Willemsma

Exactly. A lot of the NFTs have content hashes. If someone stops serving that file or the network, or others — — even gateway, a lot of people use gateways instead, which is even more fragile — — then your NFTs, at least the content behind it, go away. But having that defined relationship between Filecoin and IPFS and having that guarantee that…

Charles Cao

But with the FEVM, we don’t really have to rely on other block cross-chain solutions. We can have FEVM native solutions smart contract. And it was in one system, which means, in essence, it’s much easier than the cross-chain solutions.

Jacob Willemsma

Definitely. Very cool. Charles, thank you so much for your time and your presentation on the Lagrange DAO. And looking forward to having you back on the show at some point. Feel free to join our Discord and chat with people as they build over the weekend. Thanks for being here.

Charles Cao

Thank you.

About FilSwan

FilSwan is incubated by Binance Labs, a member of Start-up-with-Chainlink and a spaceship in the Filecoin Orbit program. FilSwan is a layer-3 cross-chain solution that helps Web3 projects to integrate data, computing, bandwidth, and payment into one suite.

Contact Us

Email| Discord | Telegram | Twitter | GitHub | Website | YouTube