Filecoin Bacalhau - Compute over Data, 28 Feb 2023

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Build With Bacalhau: Calling Bacalhau from FEVM smart contracts with Project Lilypad

Description

The aim of Project Lilypad is to create an integration for users to call Bacalhau jobs in their solidity smart contracts and hence enable interactions and innovations between on-chain and off-chain compute.

This video takes a first look at running a Stable Diffusion job and a verifiable random number function (via drand) directly from a smart contract on #FEVM

Lilypad Github: https://github.com/bacalhau-project/lilypad

============================================================
About Bacalhau:

Currently Filecoin is known for being a decentralised storage and retrieval network, providing over 17 EiB of distributed data storage capacity to the world through thousands of Storage providers globally.

We can think of this distributed storage layer of Filecoin as the Layer 0, while the ability to perform on-chain computation - or computation over state with FVM brings Layer 1 capabilities to the Filecoin stack.
Bacalhau then brings the additional capability for off-chain verifiable compute loads on large data to the Filecoin Network.. Making for some powerful tools enabling new and innovative applications to arise.

Bacalhau is a platform for faster, cost efficient, and more secure distributed computation, by running jobs next to where the data is generated and stored. Bacalhau enables users to run arbitrary Docker containers and WebAssembly (WASM) images as tasks, meaning little (or no) rewriting of the workflows that already work.

Why Bacalhau?
⚡ Process jobs fast: Jobs in Bacalhau are processed where the data was created and all jobs are parallel by default.
💰 Low cost: You get to save on ingress/egress fees you encounter whenever you move data around since jobs are processed where the data is created, as well as reusing idle computation (where available)
🔒 Secure: Data is not collected in a central location before processing, meaning all scrubbing and security can be applied at the point of collection, and a far more granular, code-based permission model.

What will you build with Bacalhau??

Links:
🌐 See more examples on the docs: https://docs.bacalhau.org
🦄 Follow our twitter @BachalauProject
👨‍👩‍👦‍👦 Join the chat: #bacalhau on the Filecoin Slack Channel (visit filecoin.io to join)

A

Right then, Ali Pacquiao show us what you think: everybody um yeah all righty I'll share. My screen. We've got a working demo as well, um but hello, everyone good to be here again as well. uh I'm Ali I've been working with the back. We are team and I. Remember, see you from uh early fbm days as well, so it's good to catch up with some old friends and see how much is getting built here but vacayo at the moment.

A

um What we're doing I've got way too many slides but I'm just going to quickly flip through them. So let me introduce back to you first, it's basically a network of open compute resources that can serve kind of any data processing workload. So that means you can run Docker containers or wasm images as tasks against data stored like on ipfs, for example, and it operates as a peer-to-peer kind of network of nodes. You can dig deeper into this by going to our docs and uh checking out all our examples and and running through them.

A

If you want to dig deeper into the architecture and how exactly it works, but I'm unlimited to time frame so I'm just going to skip through that, what we're trying do here at project Lily path, though at the moment, is combined, bacliao and fbm, because they really are better together. So um you might have seen this kind of um stack of uh what fbm allows on top of the filecoin network already um and I think we probably you're all fbm developers here.

A

So you all really know that, uh what's what it's allowing is compute over data on chain. uh So what bakeria can kind of offer? Is this computation over massive data sets of chain and we're kind of hoping to bridge the gap here uh in you know, they're not being an oracle but also provide more standards than that for what you can do with this projector Lily Pad. So uh this is just an example of some things: I'll skip over what they can do. I'll send out these slides as well to the whole group.

A

So you can all have a look at them uh in future and like read through better than I am right now, but basically project lilypad is kind of this interface or this bridge between fem and the power of bakliar computation jobs. So this means you can call back. We are jobs in your smart contracts um and to do that I'm going to actually show you a demo, so this is actually deployed on um hyperspace test net.

A

At the moment uh it the architecture we've used is we've got a base contract and then you can call it from an uh inherited contract. So you can create any sort of example, uh with a generic baklao interface, and we've got a couple of examples up on our GitHub already. So two of our current examples that we've got going at the moment are a stable diffusion job, so you can call a stable diffusion job from your smart contract and it will return uh the image that Sable diffusion image from a text prompt.

A

That's what stable diffusion is. Sorry in case you didn't know, there's a heaps of documentation out there on there that as well, that you can find out more and we've also got a dram caller. So this is kind of like operating like a verifiable random number caller on um the chain, as well, so I'm going to start by calling the stable diffusion job uh just because it takes a little longer to run an AI job. So I'm going to just call this stable diffusion job I'll give you the GitHub repo.

A

If you want to have a look through how these all work properly properly. But given the time limits, I won't go through the whole thing, so I'm just going to call this stable diffusion job with a text prompt of rainbow unicorn and it's going to go away.

A

um Call my base contract, so I have to pay some gas fees to do that. But it's going to call my base contract, which is called lilypad events. This ensures that baccodile can listen to all jobs coming in on the network and then run the jobs on the nodes in the backyard Network.

A

Once it runs that job, it will either return to the initial caller contract, whether that was an error or whether that job was fulfilled, so hopefully that job gets fulfilled and while that's running I will also use the dran caller and just just run that one at them at the moment as well. So I'll just close up my stable diffusion, One open up dran and you can get a random number between X and Y I'm.

A

Just going to run that, and at the moment you can see that the random number is sorry I'll just go back here is 281, so we'll be able to someone. Remember that we'll be able to prove that that's changed in a minute um yeah. So basically anyone can use this. Anyone just has to import our lilypad events. Contract number into their contract, which you can kind of see at the top of this contract here uh Bridge equals lilypad events with the contract number. um This is an upgradable contract, so it should.

A

It should stay this uh contract number for everyone, the hyperspace test network, but we will make sure to update you if that's not the case, so go to back at our project. Lilypad, GitHub and you'll be able to see all our changes there um and we also do have an open uh kind of notion document on use cases, value propositions, some of our goals. We'd also love to hear from you guys.

A

um If, if this is something of interest to you, if you have a specific use case, you have in mind for this we'd love to do like way more examples. uh We think this could be pretty powerful or maybe a lot of fbm uh projects that are out there, though, like it could potentially operate as I actually have slides on this. So if that's inside I'll just go back one, so yeah there's some potential use cases here. We can kind of see this with uh deal making so filecoin, IPS and storage um examples.

A

We could like, as a low-level example, use this as a ipfs interpreter or you could or like Pinter the ipfs change directly from a smart contract. Things like that could happen with this uh backyard. Bridge called Little, Lily Pad. It could operate as an oracle service. Now, potentially, this could offer price speeds. I have because we haven't tried this when we're on the kind of latency there. So it might depend on your project how up-to-date that information needs to be.

A

If that is going to work or not, though I'd love someone to try it and tell us what is and isn't working for them. uh You can also make HD pair calls, so this could operate similarly to chain links any API and that's exactly what's happening on the dram call. It's making a HTTP call to get a dramed, a random, verifiable, random number and then return that information through a script, and actually it looks like it's returned here, so I'll check the return number and it's 801, so it changed yay it working.

A

We were doing all sorts of like last minute uh fixes before we came on here, so we're not sure if it's going to work sometimes um so. This stable diffusion caller also looks like it's returned now, and we can just check that by you know calling the latest image here, just because this is the demo here's one I prepared earlier. This is kind of what gets returned. It's an ipf scid of this image, which I'm just interpreting through a HTTP Gateway at the moment.

A

um For some reason, bakuyal jobs always return rainbow unicorns with two horns, so I'm just going to go with this. This is our spirit. Animal. At this point it's a Duo chord, um so yeah I think that's probably about it. For the demo, like I said you can try this out yourself. uh Any question hit us up in our back. We are Channel, or you know, find me on slack at um Alice in Wonderland, I'm called on slack as well.

A

There's heaps of possibilities, probably more that you could share with me, we'd love to hear um your ideas for what we could do with uh this bridge as well, and what examples or use cases you'd like to see us build out. First, I'd really love to get an idea of like the road map that you'd you'd be interested in, and yeah I'm open to any questions as well. Okay, I think. Maybe we have a couple in the chat. Yeah enjoy, calling that's right Simon um by the way.

A

I also did this in combination with Simon Worthington who's here tonight, as well. He's done a lot of work on the backyard side of this. So thank you so much Simon.

A

um So at the moment it's brief that man there's no monetization on the back of your nodes that will change likely in the future, but at the moment everything's we're running it for free, it's basically probably good at the moment, so uh go for your life. um Yeah I'll send the link to this in the chat.

A

Also, so here's a link to the pub uh and also to uh the notion which you can get to by going to bitly as well as bitly, lilypad Dash project or project Lily Pad, one of the other yeah, something like that project. Lily Pad there it is oh yeah, yeah and I'll also be in Denver. So if anyone is around and wants to chat to me about this project come and hit me up at the hacker Pace I'll be there on the 28th I'm, not sure about the first.

A

Yet I might have some competing things happening, but I'd love to see some of your demos, so I'll try and make it um and I've also got another demo kind of related to this, which is AI art nfts on fbm, um and that is with the back of the Hour project as well. So there's more, you can see there I think I've gone away over five minutes. Sorry I will shut up now, uh yeah happy to answer any other questions as well. Thanks everyone.