[Music] Hello everyone. I'm Alex Pashinski. I'm engineer at Fluence and let's do our workshop. So let's start with intro like why what's the purpose of fluence? Like everyone uses clouds.
Clouds are everywhere but they have like really really huge problems in terms of web free also. It's vendor locking. If you do like whole your business on top of like AWS and you like been banned, you're basically and uh you need to spend a lot of money and resources to just migrate and it's really hard. It's noncompatible also. And uh big clouds uh charge really big costs on uh on top of the like cloud expenses.
It's like really really expensive also. Yeah, you have censorship risks and lack of transparency how your data is stored and how like everything works. So that's why we built Fluence. We are like decentralized cloud. We trying to build like the same but in web freeway and with like competitive and uh enterprise great infrastructure.
So you don't rely on just exact one data center or exact one cloud provider and you can choose like using the same interface uh between like multiple data centers. So uh our key features is like we are really cheap compared to every other cloud and there's no vendor lock in. You just can choose any location. You can choose what hardware do you want. If you are business and come for us we can arrange uh any hardware with our providers to like tail your needs like SJX GPUs and stuff.
Uh also we have multiple locations and for now we have virtual servers but on our road map we have containers, GPUs, managed services and everything any cloud has. So for bounties for bounties we have like two types of them. First is our bounties and second is a joint ones. So for joint ones uh we partnered with OG zeroG laps and one inch. So if you do any bounty from this projects but for back end you use fluence you deploy a back end where you can receive uh price from us too.
Uh and for our bounties we have uh best use of fluence virtual services in AI. Uh you can do your project but keep in mind that uh for now fluence is only CPU cloud. So you can come up with idea how to use like small models for your needs and uh like don't expect to for CPU to like work as GPUs with AI and uh the second one is reusable uh virtual images virtual server images. So for now on fluence you can deploy like standard images like Ubuntu or Debian or any other but you can build like with Nix O or any other tool uh custom image that can be uh have pre-installed all have pre-installed tools like for example GitHub runner so our users of Fence platform can just choose your image and deploy it and have like full-blown GitHub actions actor that works and for third one is kind of obscure. You can look how fluence chain is uh working our rollup and uh propose how to do multi- chain staking because for now uh we have like our token on Ethereum actually and uh we have we have a roll up on arbitum orbit which typed for Ethereum and to stake you need to just buy token on unis swap and ethereum then bridge it to fluence and then stake and it's kind of like a lot of hustle but if you can come up with some idea to optimize it you and get a bounty.
So, let's have a demo cuz I like demos most. Uh, so that's how our console looks like. It just like most uh most uh web two clouds. You can choose any location you like. Uh for now, we have like three in available capacity.
You can choose instances that ties for your need. It's basic configuration but it can be like uh have extended storage and uh also you like name your instance. You can like choose what ports to open and what our ports will be closed. You can choose like images and uh you can choose custom image if you like do a bounty for custom image. uh you can have like SHK are the new ones and uh let me make it cheaper and then you can deploy it and uh soon it will be like uh set up for you and uh you can find it in this tab.
So let's look how it looks when spin up. So I deployed this image just before workshop and now you have like VM ID you have a location you have all certification that data center has also you have your uh your IP and uh to make a s access and do everything you want with your uh with your VM and you can also delete it and terminate and like it's like basic way to use web UI for to deploy to fluence Also uh you can use API it's uh really good for your like project that utilizes it and uh first you can check documentation on fluence.dev all links available on if global website so it just explain all the like things we renting and uh let's look at the APIs. So AP API is kind of like simple but it allows you to deploy all the stuff look for all the offers and uh get the statuses of your VMs. So let's create a pack and uh deploy something via API.
So let's create a K and uh let's deploy it. So I have prepared comments because I'm too lazy to copy paste it. So let's put our K. Oh, and it's deployed. And we should check it here.
So this VM uh Belgium that I deployed in UI and this one I deployed in V API. Also you can deploy multiple VMs at once. You just uh need to specify like where is what? That's one instance. You can specify like two, four, five instances and yeah and it should like spin up in couple of seconds but it depends on location.
So for hackathon you can get free credits for our platform and uh and use it to just try it try it out. So first thing you fill a form by this link uh and login with the same email or GitHub that you have and we approve you. You wait for confirmation. We approve you again the credits then you regain and feel free to explore the platform and u also if you have any question uh about like uh if you come up with some idea and you want to validate it come to boo or just write in discord but better come to boo and uh I I will speak to you and explain it's possible or not and uh yeah we have also support via discord we have a bunch of a available online. Uh but basically it if you have like any questions I can show it or answer.
Yeah.
So if you
Hi. Um so it's one thing to be banned from AWS because you're in the wrong country.
Yeah. But what if you do something like truly malicious with one of these V VMs and just send out a load of spam or something like that? What happens then?
So, okay, I got it. So how it happens right now uh like it works like this uh providers also have some like forms and uh like local government or local providers uh send notification for providers who are like list of IPs uh that something wrong is happening on where IPs. So after that they send notification for us and we suspend VMs and like trying to understand with snapshots and stuff what's happening. So for now it's like we receive notification when we receive notification when client receive notification. So it's like kind of case by case.
Yeah. So it's like like
um is it possible to uh deploy something that is dynamically allocated in terms of like where it's deployed? Uh, so that I noticed you have
I noticed you have like um Poland and uh a couple other location.
So like you're asking about resizing image like resizing instances. Yeah. Or what?
Well, like depending on where the traffic is coming from, is it possible to have that dynamically move to a different data center and then have it um have it be
like move your volumes and like all the stuff from one uh VM to another or what? Yeah, I'm trying to do like um I'm thinking like basically a um like a multi-data center um traffic load so that if a data if a request for whatever it is that I'm deploying comes in from Poland then um and that hits a certain threshold then it should deploy a new instance in in Poland. Does that exist on your influence or is that like something that I would have to do with the API? So like migr migration of VMs right now is like not implemented. We almost finished with like uh inside data center migration um and like migration between different providers is on our road map but right now we have like VMs that you deploy them if you want to migrate you can do it like manually just deploy second VM copy all the stuff uh from one to like to the second and yeah so it's kind of like really basic level of cloud you have VMs but you can do anything but not all like things things out of the box.
Yeah. So,
I was just wondering um in terms of the credits uh how how much can we interact with the servers for those credits to be just
Can you speak louder please? I just don't hear.
So, I was just wondering if the credits that uh you are giving us are sufficient for like playing around. I don't know heavy things.
Yeah, good question. So, okay, you will given $20 credits and if you just look, you can rent like 32 vCPU instance and for a day it's $4. So, it's like 5 days of VM. If you like really really need for your projects more credits, you can come to us and we will arrange it. But it should be justified, but it's kind of enough.
Yeah,
thank you. So like smaller instances like less than a dollar a day. So yeah, it should be enough.
Okay. So if you have any question later, just come up to the booth and we are just here. So have a good day, have a good hackathon. [Music]
Automatic transcript — names and jargon may be misspelled.