From c3a0322884f01c3c18d4aa9d205167fbcb0b1724 Mon Sep 17 00:00:00 2001 From: Drinkwater Date: Wed, 4 Oct 2023 19:11:32 -0400 Subject: [PATCH] sig-providers Create 009-2023-09-27.md (#313) * sig-providers Create 009-2023-09-27.md Signed-off-by: Drinkwater * Update 009-2023-09-27.md add recording Signed-off-by: Drinkwater --------- Signed-off-by: Drinkwater --- sig-providers/meetings/009-2023-09-27.md | 331 +++++++++++++++++++++++ 1 file changed, 331 insertions(+) create mode 100644 sig-providers/meetings/009-2023-09-27.md diff --git a/sig-providers/meetings/009-2023-09-27.md b/sig-providers/meetings/009-2023-09-27.md new file mode 100644 index 00000000..3d196804 --- /dev/null +++ b/sig-providers/meetings/009-2023-09-27.md @@ -0,0 +1,331 @@ + + # Akash Network - Providers Special Interest Group (SIG) - Meeting #9 + +## Agenda + +- Follow up discussion on GPU Provider Setup. +- Update from Praetor on all of the work that they have been doing. +- Open Discussion on all things related to Akash Providers. + + +## Meeting Details + +- Date: Wednesday, September 27th, 2023 +- Time: 08:00 AM PT (Pacific Time) +- [Recording](https://c5g4xir7o2v6il2syhpnqgaooslncply3whcz64ykkn4anqppzha.arweave.net/F03Loj92q-QvUsHe2BgOdJbRPXjdjiz7mFKbwDYPfk4) +- [Transcript](#transcript) + +## Participants + +- Andrew Mello +- Andrey Arapov +- Anil Murty +- Damir Simpovic +- Deathless +- Deval Patel +- Jigar Patel +- Max +- Rodrigo Rochin +- Scott Carruthers +- Tyler Wright +- Zach Ross +- Zeke E + + +## Notes + + +- Scott mentioned that the documentation for Provider build experience will be consolidated with GPUs and CPUs. This information will be shared on Discord for feedback. Link can be found [here](https://docs.akash.network/providers/build-a-cloud-provider) +- Praetor team gave an update on the work they have been doing. +- Andrew Mello shared the work he was doing with AkashDash, AkashOS, and around content morderation Chapperone tool. + +Praetor team: + +* Issues with persistent storage - worked on this month +* Persistent storage was not available for single node or GPUs. This has been fixed. +* All providers should be helm. Praetor team can help anyone that needs to move to helm from the old format. +* Working on moderation API - should be done in two weeks time. +* Management API - providers should see the kinds of leases on their own provider, and they should be able to close the lease if they want. +* Some requirements and docs here on content moderation can be found here [https://github.com/akash-network/community/tree/main/wg-content-moderation](https://github.com/akash-network/community/tree/main/wg-content-moderation) +* Moderation API will provide filters for kinds of workloads. +* 3 filters: + * (address - white listing) + * (Host name) + * (Images) +* Zach Ross wants to get involved in the implementation with the Praetor team. +* Praetor wants to give a demo at the next sig meeting. +* Praetor team Will distribute the binaries to those that want to help test. +* Scott asked are there any thoughts on whats next after version 1 coming soon? +* Praetor team wants to implement USDC on Providers. + +Andrew Mello updated his Chaperone tool and other work he has been doing. +* Andrew Mello should have something ready by the end of the week to share with the community. +* Andrew mello built Chaperone for the provider. +* DMCA incidents + - Andrew mello had some ebook sharing + - Movies were being shared Andrew’s provider +* AkashDash tool is being used. There are many checks happening and the process is automated. +* AkashOS is going to be Andrew Mello’s AKash-a-thon submission. He will share with the community after the submission. + - [https://akashdash.com/](https://akashdash.com/) + - [https://akashos.com](https://akashos.com/) + - [https://github.com/cryptoandcoffee/akashos](https://github.com/cryptoandcoffee/akashos) + + +### Action Items + +- Zach Ross to work with Praetor team on content moderation +- Praetor to share demo of content moderation work at the next sig-providers meeting + +# **Transcript** + +_This editable transcript was computer generated and might contain errors. People can also change the text after it was created._ + +Tyler Wright: All right. Welcome everybody. To providers monthly meeting number eight or nine. I believe number nine. again During this special interest group for the providers, participants talk about all things related to the provider on the cost. there's been again, If you providers GPU providers, they're provider support tools, like, pretor, and others. There's documentation related to just provider experience. So anything is free game. As long as it relates to the provider on Akash, + +Tyler Wright: out of the last monthly meeting, we got an update from the page war team and all the work that they've been doing since the upgrade has happened again, Shout out to Damir who's on this call, who has been supporting and helping out to pray to our team and testing. A lot of the technology that again the paid protein has been implementing including just one click in a couple clicks to become a CPU Multi-note provider. So again, shout out to the prey to our team and simply for all their efforts, getting folks ready to Again, use their GUI tool to become a provider on the Akash network. + +Tyler Wright: There's also been some documentation that's been updated and continues to be updated by Andre and Scott from the Overclock Labs Core team. Scott just mentioned it during six support, but again, there'll be some documentation coming down the pipe that basically is a consolidation of the provider setup experience to this point, I think there's been a provider setup experience for CPUs and then another provider setup experience as it relates to GPUs, I know the Scott and Andre I've been working together to consolidate that into one X provider set up document that is encompasses all kinds of providers again whether GPUs or CPUs. I think, Scott has tested that documentation internally and is going to start to share with folks in the community and publicly to start tests. And again, it can be improved on if it needs to be improved on. + +Tyler Wright: No, I'm Scott. Do you have anything that you want to add there before we continue on with the agenda? + +Scott Carruthers: Yeah, no, that's basically yes. So yeah. Probably in the next hour, I'll announce this and the Providers announcement channel. But yeah, just as Tyler stated, going back to Testnet. And then right after we did, they may not upgrade with the launch of GPUs. We had separate documentation to become supper documentation to bring you all the way through the Kubernetes Cluster setup into the provider building. And via driver and runtime installations. So obviously, we don't want to maintain disparate documentation over time if we need to. Update something in one document we have to remember to update in any other. So obviously we want all that consolidated so now they GPU provider build experience will. + +Scott Carruthers: Be somewhat synonymous with persistent storage or IP leases, where you would go through the entire build of your Kubernetes cluster and a cost provider build. And then, after that is complete, there's a optional step to enable GPU resources. So all that will make sense when I release documentation. But again, it'll be great. To just have one consolidated set of instructions going forward. + +Tyler Wright: Thank you. one quick thing to note, just for everybody's participating. I know there's some folks on the court on the Overclock Labs Core team, that always attend providers meeting amongst many others. Just a quick note just for today, There might be a chance that some folks have to job drop, just because there is a validator upgrade that is planned for about an hour from now. Again, everything seems to be in line right now. There's been contacted and they're just getting themselves ready for this upgrade, but just want to call out that if an issue does come up over the next hour, on some, folks from the core team may have to drop off to just handle that issue. So I just want to make everybody aware of that just in the same page. + +Tyler Wright: Thank you, Scott, for that update on documentation again. I know that as folks in the provider SIG channel and the provider discord channel, and have questions about documentation or Andre and Scott Joe, a great job of helping support folks, and those channels. And then again, if anything needs to be updated in the documentation, they make sure to update it. If anybody here has a passion or desire to help support documentation, please feel free to reach out to me. I think I put a message in SIG documentation as well. We're always looking for more folks that are technical in the community. + +00:05:00 + +Tyler Wright: that might want to help support documentation because again the Overclock Lab Core team is very small and doing a number of things. So the more folks that can help out with documentation, even if it's just, Phase One and Phase Two and then somebody else from again SIG documentation or this group, can edit it. That would be much improved than again. Just not having those extra hands to help with moving along. I know that Jigger and Devala here. I want to give them an opportunity to just provide a quick update on the work that they've been doing around the Pray Toward tool. And then after that again, update on the work that they've been doing around the content, moderation efforts related to the Working group Content, moderation, So without further ado, I'll just hand it over to Jigarette about to get us to give us an update. All the work that they've been doing + +Jigar Patel: Thanks, Tyler. So Let me just. + +Jigar Patel: So our team has been working tirelessly to support providers in the network. we've been doing a multiple things, as you can see that there have been issues with the personal storage for single or not providers. That was the major thing that we did this month. There was a demand that person was not available for ngpus or single provider. So we added that we also committed the price scripting, which had a major change. Also have a lot of providers to become a cloud provider with the GPUs on Akash. + +Jigar Patel: also, I think most of the service based provider we help them to move into helm. So, I think we are now aligned with the Akash. So, if anybody still in the service provider and request for any help, please just messages. I think we will be able to help them to move into helmets. It's very straightforward process, was Just we need to be there. So Something happens. We can just rectify it right away. + +Jigar Patel: We also started working on the moderation API as well. And it's probably we will finish In two weeks of time and also as a Tara asked me to just explain what we're doing in the Moderation and Management API. So we started this working group in starting of this year to solve the issue The providers who are currently in the network has issue, where one unknown people, + +Jigar Patel: Put a lot of things on their providers and that might not be appropriate for that country or the laws where they're predator is heading, So we decided to implement management API and moderation EPA. So I'll just briefly explain. I think Jack Ross is on the call. So I just wanted to explain what we're doing in management API and what we're doing in moderation API. so, in a management API, our goal is to Provide APIs where + +Jigar Patel: Providers, who can list the leases from their own providers. So if I'm a provider them for tan loses is running on my provider, I should able to see, what kind of a workloads? They are running on my providers, right? One particular leases and I should able to close that list. If something appropriate I found it right plus that L you're able to so that they can see if it's a front end application, they will be able to see what kind of + +Jigar Patel: Things, they are running, So, if something is not appropriate, they can close it. So that's the management API Just like The other side of coin is the moderation API because we have to Also look into the moderation where from the day one, they should not able to put a lease on the provider, if we put some filter on it. So + +00:10:00 + +Jigar Patel: In a moderation FBA, we are basically adding filters. So currently We basically decided to add three filters so address. So that's basically white listing, right? So that's also add into the ministics upgrade. But two other main filters that we are adding is if certain words in the host name, you can just, not host there. All this on your providers and images. So let's say + +Jigar Patel: Approved immediate list is basically. You say, what, I wanted to only host approved images from this list so they can do this right? We are adding a list API filter ad so you can add filters, you can remove the filter. And basically when you do the order or we needed the bead. Basically this process, we'll call this filter API and check if that filters, if it passes, then it returned to and then it providers will be on that order otherwise you will not + +Jigar Patel: So, that's basically, the general idea that what we are doing. If you have any question, please let me know. So I can clarify more in detail but I was in this moderation and management APA. + +Tyler Wright: Thank I believe Also, put a speck that came out of those working group sessions in the chat for people to also reference if they need to Go ahead. Zach is here unmuted. You have any questions? + +Zach Ross: I was going to say it all looks good. I'm just curious about more specifically how to get involved in the implementation part of it just to get Kind of Digging into the next gear. + +Jigar Patel: Absolutely, so Zach we can create a one discord group and we can corrobat on that again, We started working on it. So if I have to talk to my team, if they need any help, basically, we can just involve in you into the process and we can just go ahead and like to combine work basically. + +Zach Ross: Yeah, it sounds great. + +Tyler Wright: Jigar, I know that you mentioned about two weeks out, I believe on the moderation API, How far do you feel like you are with a solution that might be, at least tested with some providers in the community and then more widely scared. + +Jigar Patel: Yeah, let the world speak on it. + +Tyler Wright: Okay. + +Deval Patel: Yeah, so we plan intend to give demo of first in the next sign by the next segmenting, right before that will ask Damir or other people we'll distribute the binaries and try to, see if that works fine, So, that kind of thing we are planning to do And keep a tab on, things with you and other people in the Crater private channel, So, yeah, I will let you know, but our goal is to have everything ready by next provider significantly. + +Tyler Wright: Take a knee, I'll just put another great point in the chat again. I'm fine to keep some of these conversations that have sensitive information as private, but as much in this conversation that we can keep in the public setting especially of Zach, I'm gonna get involved in supporting implementation with the prey to our team. We do have that working group content moderation channel where we can talk about updates. Get people involved for testing over our next month, they sync and just make sure that we're continuing to track because I know this is something that again, we talked about the last steering committee as a high priority item, and I know that everybody in the communities and thinking about this and working on it. Go ahead Scott. + +00:15:00 + +Scott Carruthers: Yeah, I was just wondering Jigarh and Deval you mentioned whitelist in. Hostname restrictions and in approved images. I'm just wondering. So those are obviously, tremendous starts and that's going to get A lot of value that's not there today. So, yeah, obviously, tremendous start. I'm just wondering if and I know going back to our content moderation working group sessions early, in the year, there were some discussion around this. I'm just wondering once you implement that initial version of the API, do you have any thoughts on what's next? And if you don't have that at this point, that's fine. But I just wondering after you implement white listing, host names and approved. Images Do you have any thing in mind? What? Possibly might be introduced into that After that. + +Deval Patel: So I think Jigar mentioned about the management API, which is part of, the overall content management in terms of let them manage a couple of things in terms of by their own, right? So there will be how they'll look in the content of the leases and deployments, So that's one thing, right? Other than that, nothing came out of the working group content moderation. Right. But, if needed, and what we can do is that once we, implement all these things, which actually finalized in the work in the working group of content moderation. All right, and things, right? the we can revisit that group and see whether the areas of improvements are right and we can work on that. How much? + +Scott Carruthers: Okay, sounds good. Yeah, thanks. + +Deval Patel: so, Okay, just a quick point after the content moderation, things like we were planning to do actually work on with the core team and things. If the + +Deval Patel: USB c and stable payments are available. We would like to integrate that, that's the next things we wanted to do, but it's up to community what they decide in things to them can align with them. + +Tyler Wright: Excellent, thank you for the call. it's great to know what you all. I think about next even beyond content, moderation. Go ahead, Andrew. + +Andrew Mello: Hi everyone, can you hear me? Okay, great. Thanks, Jigar for the update and Deval for the update on the moderation, there, I was. Since our last meeting able to roll out the chaperone solution, that was up for discussion over the last few months, that has been now turned into a Python app that runs existed on the basically control or masternode that has cuddle or cube config access. + +Andrew Mello: And that Python app will actively close and moderate. Any deployments not only based on image name and, kind of the high level characteristics, but it will actually go in and do scans for keywords. E checks for VPNs, it will do process each checks for tour and sockets. Mainly the things that will basically get you banned from an ISP, in a weekend if they're not solved for immediately and that has now live and implemented on Bdl DOT computer and SFO Dot computer. So if you want to try that yourself and see what the user experience is like you can go ahead and do a test deployment of bitcoin on Bdl computer or SFO dot computer. And you should see your deployment closed within about 60 seconds or so depending on + +Andrew Mello: Where the scan is at the time. But the scan runs on a continuous loop, basically scans. Every single Pod excluding the system pods and we'll do that, forever. So I did implement and finally move forward with that because as I mentioned, during our last calls, DMCA complaints were coming in, It was time to Actually, get a solution developed here and I was waiting to roll that out at the end of the week here, when that's gonna be packaged up in a car show us, which is going to be my HACKATHON submission project. So all of that Python work should be available in the public by the end of the week. And if anyone's running a provider that actively needs content, moderation, we have a solution out there That's running and working and mainly right now it's focused on torrent VPNs and people trying to manipulate connections to do things that otherwise they shouldn't be doing in my experience. I haven't really seen anyone using malicious + +00:20:00 + +Andrew Mello: Names that, if anything they use just a Ubuntu, 22, image, and then do weird things inside of it. And then I'm just my final note and feedback to the Prater team would be to really, in the best of way possible. Avoid any interaction between the user person running the provider and that user data, the way that chaperone the Python app is designed, is that, even, the administrator isn't gonna see anything, unless it's absolutely flagged or caught and pulled out. And all they see is a file name or a location. They're not actually gonna see the content. So with that in mind, just some ideas and some heads up. Moving forward for you guys. Yep. Develop yeah. + +Deval Patel: Yeah, so it's definitely. We're gonna try it out, And are very like where we can use it or our utilize that Definitely. Another thing you mentioned about the user data and touching on that, we do not encourage the providers to touch down on the user data but they do have already control of the capital Keep CDL by their own, right? So they by default,… + +Andrew Mello: But yeah. + +Deval Patel: They have the daily access of that, We just want to provide them in a more, user-friendly way of looking at things, we don't want them to go inside the application and see the things are, like that right. We definitely do not encourage that, It's just for them to know. What kind of application they are running. That's + +Andrew Mello: So my solution as I noted was if a match is found of suspicious content. simply the Location is Found deployment closed. There is no, displaying of what the content of that file was it was a match enough to hit on the programmatic search, right? It's kind of the way that I've handled it and as well. Then we also understand of course, providers have, full access to an entire Kubernetes cluster, the administrator, they're gonna have access to that data. then there's the other side of it of determining, a providers terms of service and the way I kind of have tried to model my providers and my terms of service is around, just like a product, right? Digital ocean. Although they have the right to go into my data, It's in the terms of service that they're only going to scan my data or check, that I'm not in breach of + +Andrew Mello: Terms of service, they're not going to go in there, poking around reading my files. So it's that similar kind of fine line but we can definitely design solutions and it's clear now that solutions can be designed that do not have to, me as the administrator interact with that content or data outside of knowing the location of a file and that, the deployment was pretty close. So I'm the tickets already gone by that point. Yeah. Jigar. Yeah. + +Jigar Patel: Andrew, that's great. you created this but actually if we go back to our content modulation discussion, + +Jigar Patel: The basic point was to shift the liability and not to run anything automated. Because then it's if something missed up, if our application or your application is miss some content and if not able to detect something malicious happening and will not able to remove it, Then the problem comes arise of the liability, Who's liability was that, right? It's a Caribbean or is it the providers liability, right? So that's why the one that we are creating basically will be saved with a caution network, Or the library. That means that we are just giving tools to the providers. what you can manage this, It's your liability. And That's the shift that we discuss the Scott also. raise that question that everybody and the + +Jigar Patel: address and calls quite + +Andrew Mello: So just my feedback on that is in the real world in being an actual operator of a provider that now has DMC complaints from my ISP. There is no way that this could not be fully automated I cannot sleep at night + +00:25:00 + +Andrew Mello: Waking up having to check a cluster every single morning or I cannot be away from my computer for a week and then come back and find out that I have four dmca complaints because somebody wanted to try to rent for two hours on the way that chaperone is Number one is to protect the provider. If the provider is, blocked by their ISP, the black provider gets into legal, trouble because they are hosting malicious content or a virus or a botnet. We've got bigger problems Then the deployer complaining that their deployment went down. The bigger problem is they were running a legal malicious nefarious content. + +Andrew Mello: And that was dealt with immediately. And again, I think that spreading the word and making it clear that a caution is not a place to do that stuff. And if you do that stuff, yeah, you should probably expect that you're, torrent deployment might just disappear from select providers, that are trying to create a stabilized network where they're not putting the network connectivity at risk. Just to be, open on that note again, running the largest provider, it is in my absolute best interest to run the most amount of deployments. And the way I do that now, is by activating chaperone on my provider instances, and making sure that I can sleep at night knowing that not one guy is gonna take down the whole network or the entire ISP connection. So with that, I think the real World Implications number one, have of course, to protect the providers infrastructure, their network. + +Andrew Mello: Oceans their subscription services to make the provider even exist first. And then of course we need to good policy acceptable, use policy terms of conditions and people need to understand when they deployed a certain providers that certain conditions need to be met and I've written now in my conditions, in those terms of services, for those specific providers, no legal content, no torrenting, no VPNs. And just making that very clear and The providers are still happily running and now I can sleep at night because again, it has to be automated if it's not automated. There is no way that You cannot scale, a provider or the network. + +Jigar Patel: Yeah, and I agree with you, just the line that we are drawing. Here didn't optional tool that we are providing, So your tool is gonna be optional. It's not mandatory that everybody has to put into it, And what we're doing is basically, we are sipping, this APIs with the library, that's the two part of the basically broader management or sorry, moderation, that we are doing in the network if you okay with that. + +Andrew Mello: I understand and I hope you understand my, real world operational set of problems that have to be resolved and have been now resolved and implemented through this python. moderation just directly talking to the Kubernetes cluster and interacting with it in the most automated way and a way that keeps network performance and everybody happy on the network and not as getting shut down because somebody wants to point + +Jigar Patel: Basically, these things can be provider. + +Andrew Mello: Crap. And anybody used to enable it or… + +Jigar Patel: Yeah. provider has a right to install? Yeah. + +Andrew Mello: disable it or change the terms, the keywords are up for you to decide what keywords you want to moderate. I don't care. again it's a tool that's built for providers to add to their provider if they choose, it's not required. Nope. + +Tyler Wright: Go ahead and kneel and then rodr. + +Anil Murty: Yeah, so I just want to jump in here rew. So thanks for making that solution, definitely going to be really useful. I just one question in one comment so the comment first is sort of I agree with what Jigar is saying which is why somebody should be. There's no reason why Your solution or the one that Jigar is working on is the only solution or the right solution. It's entirely up to the provider, if a provider feels like they want,… + +Andrew Mello: Yeah. + +Anil Murty: you have this run automated and be okay with the potential risk of losing Deployments as a result of it sometimes killing Deployments, that's entirely up to the provider. So I don't think we need to argue about which solution is better. We definitely can use more than one is. + +00:30:00 + +Andrew Mello: Yeah, I just wanted to share the big updates since we all know. + +Anil Murty: Yeah, sure. + +Andrew Mello: Yeah, not arguing at all. they're working on their own thing. I'm trying to solve for the now and reality of running the largest provider on the network. And so,… + +Anil Murty: 100. + +Andrew Mello: we now have a tool to handle the situation. Yep. + +Anil Murty: do you have that open source somewhere that we can refer? + +Andrew Mello: Yeah, I noted at the beginning, that's part of my HACKATHON submission,… + +Anil Murty: Okay. + +Andrew Mello: which will be at the end of the week. I'll share that have URL at the end of the week when everything's live and packaged up for the submission for the HACKATHON. Yeah. + +Anil Murty: Amazing. And then I did look at your TOS and AOP submissions as well. My only feedback on that was if we can call that a reference as opposed to this is the terms of those because I mean correct me if I'm getting this wrong but that's a legal document right? So you're basically saying this is the terms of service to using my provider and I don't know say that you have written it pretty well,… + +Andrew Mello: All right, it's a bit. + +Anil Murty: but I think if it's going to be something that you put out there and tell the providers to use we should indicate that it's a reference and the providers are welcome to use Not use it modified whatever. + +Andrew Mello: Yeah, it's definitely reference if you look at it inside, everything's a placeholder. I'm sure we can change the name on it. I wasn't so specific because we were in some working group there, I imagine people understood so working group. so, + +Anil Murty: Yeah. Yeah. I mean yeah hundred percent in the working group is totally fine, but if you're going to put this in the code base, I think we just make sure people are clear that it's a reference and they should apply their own, or maybe hire a lawyer, or something and decide to + +Andrew Mello: Absolutely. It's just to get a template out there and show people what I'm working with. And yeah, + +Anil Murty: Perfect. Yeah, thank you for doing that. Appreciate it. + +Tyler Wright: Yeah so Andrew I just threw that in the chat. I think a deal gave you direct feedback. If you can update those PRS with the suggested naming then again those can be merged. Go ahead, Roger, + +Rodrigo Rochin: And I was gonna ask Andrew where I could test the Python app by they already answered that. And I like both ideas. I like Anil said, providers can choose whatever tool they want to use, but if we want to check out Andrew, + +Andrew Mello: Thanks so much. Yeah, I can sleep at night again and I don't have to wake up every morning. jumping over on lens workspace there, checking, what somebody deployed overnight. I mean, it's just, over time, if you need something to do this for you. So, + +Rodrigo Rochin: yeah, I mean I have a smoke writer and I just wanted to check it out but I understand what you mean because sometimes I knew it I don't go in and check the provider every day, I check it every I don't know two or three days but it will definitely help. + +Andrew Mello: Yeah, and that's what happened to me. these dmca hits came in and of course, the deployments weren't even there, to check or look at of the month ago when they got deployed and the ISP sending me an email, weeks later. And it's a mess. So yeah. Automated so it doesn't happen anymore. Yes, I know. + +Tyler Wright: Go ahead and go ahead and + +Anil Murty: but did you find out what triggered the DMCA folks to reach out to you and if there have been other instances of this happening with any other providers besides you, + +Andrew Mello: Yeah. Someone was sharing some electrical ebook, another one. We're a couple of movies in July, the ISP sends you a direct content ID, They basically taken a hash mark of the file size and the metadata and they see it running over your ISP over the wire at any point. They're like, you must be sharing, it was one of the oceans movies, oceans 11 or something someone and this is absolutely, attributed to the cluster. There's nothing else comes from. Yeah. + +Anil Murty: Got it. Thank you. + +Tyler Wright: Again, I appreciate this discussion obviously multiple people are working on Different solutions to help solve this problem for folks in the community. So, again much appreciate that. As next steps Andrew mentioned that his solution is going to be packaged and be a part of the acasathon which again closes this Friday. After that submission is complete. Then again, I think folks like Roger and other people in the community would love to test the Akash OS and that's a chaperone solution. I know that the pretor teams that they are going to be working over the next couple of weeks on the content on the Moderation API and the management API. And again, Zach has offered support in helping implement to again. + +00:35:00 + +Tyler Wright: Make that process more efficient and speed up that process of preparatory. Team needs it as next steps. Again, After this Friday, I'm going to check in with folks next week and maybe we can use the Content Discord channel for testing to start again. We've had great conversation over this session and previous sessions about content, moderation before we start, Pushing these solutions wide across the community and start telling providers to use these solutions. I think we need to go through a testing period. Where again, folks on this call, folks is a part of this stick providers. + +Tyler Wright: As part of this provider special interest group can test out these various solutions or maxibly skip feedback to the pre-tor team, get feedback to Andrew Mello and make sure that we're all very comfortable before. We again suggest these solutions to the wider community. So again I think that both solutions seem like they're very close. If not ready to go something Andrew Mello's solution being implemented on its own the next step before again, a wider range of sharing these solutions with the provider community is testing amongst people in this SIG insiders in vanguards. I just want to make sure that that's very clear to everybody here as the pertains to next steps. Where beyond implementation of these solutions is testing. Once a testing has been completed and signed off by again, members of this big, then we can move forward and start sharing these with the wider audience and start to again market these solutions or + +Tyler Wright: I stare with various providers. Does that make sense his next action items for folks? + +Tyler Wright: Cool one thumbs up. So I appreciate All right. is there any other conversation related to the provider outside of content moderation that anybody wants to talk about right now? Whether it be maybe a specific provider, experience, or anything related to running provider on a cash. Go ahead, Andrew. + +Andrew Mello: Yeah, so I know I took up half the call ready. Sorry about that but had a lot of updates on this end and I missed the last one. I think so. Number one are the akash dash product that I launched this summer that has been performing extremely well. we're now surpassing it looks like 35. They're 3600 checks? So that's about 20 2500, checks of provider stats in about a month and a half two months, which is really good, I'm seeing it go up every day. So it's nice to see that providers are using that and they're finding it on Google and whatnot. and it's proving to be a really Nice site for people to land on the other update I have is that, Akash is basically going to be my HACKATHON submission coming here at the end of the week, and there is now a website. + +Andrew Mello: Before that with just a little bit more information. And that will be where it lives in the future. And then probably the biggest update is, let's see here. One sec. + +Andrew Mello: Probably I'd say the biggest update here is the Github is I guess I should just share it now, because should I share it does it matter it's going on to the public on Friday? right. + +Tyler Wright: yeah, yeah, again, you feel free + +Andrew Mello: So to get up, it's coming together. And probably the most important thing that's been updated there is the Readme and right at the top defines, the architecture of how this thing is built. And it's about as simple as can be with all the complexity required for a cost. So what it does is it takes a, generic Ubuntu, 22.04 Server Auto install and then uses Cloud in it and user data scripts on top of that to create a, entire provider and Kubernetes cluster for the user and fully provision And the only thing the user has to do is answer basically for questions during the setup process and send the provider some funds and they have up and running provider. So the main features that it's really going to have I think one of the most exciting + +00:40:00 + +Andrew Mello: Months that I'm really buttoning up how this will be done but is basically zero configuration networking and DNS. So what that will mean is that if a user of a costco s is able to get through the install the software will handle all the networking and routing for them until they go ahead and update with a custom domain. So what does that mean? It means we can spin up and turn on providers without asking them to go into their firewall or change networking settings or do weird stuff to set up, DNS records which Cuts Out, Probably, 30% of people who want to become providers, that's just way too technical for them. And the unique way that, I basically set up to do that is using the Open Source tool chisel which can do full on TCP and UDP. + +Andrew Mello: Routing. So with that we might be able to do something like, Provider XYZ dot Akash os.com and basically just give all new providers A provider address default routing and they go in later and update it with a custom domain name. If they are even get to that point. I'm so the whole point here is to make the process completely first of all, secure and on premise. So this is supposed to be run. + +Andrew Mello: VMs, on bare metal, inside of your own network, on your own hardware at no time do we require SSH access. And then finally, there is a UI and a graphical user interface for the user. Once they finish, the installation process. So I figured I'd give a little bit of a touch on this today because we probably won't meet for another month. And it's going to be released here between now and then and I wanted to feel any quick questions or take any quick feedback? Or just Bring it to light here today. + +Tyler Wright: Does anyone have any initial questions? I know the people have to dive in a little bit more over time. And again once the tools been submitted folks will start testing I think as early as next week but does anyone have any questions or initial questions or comments? I could feel free to have any follow-up discussions in providers discord channel. + +Tyler Wright: There's nothing else. I just want to give everybody a quick heads up. I know that we're running up on time and have about 15 minutes left. One thing that I wanted to do today. And again, there's a steering committee happening tomorrow at 11 am Pacific. So, might continue it tomorrow, and continue and discord. + +Tyler Wright: Members of the core team and obviously members of the community. And again, this information will be made public just trying to collect some information from members of the community about the use of the SDL and the SDL experience on I know that people are deploying and providers I got on a network are having to interact with the In some way, shape or form going to share a small type form to get feedback, but if anybody has any initial feedback and again, that type one will be shared in discord for people to give again just quick feedback on their experience with SDL. + +Tyler Wright: Things that are bad with the arm. Just trying to again better understand from members of the community, some pain points around the SDL and So I know that we have a little bit of time left and everybody needs to feel free. But I just wanted to hear if anybody had an, initial thoughts about their experience with the SDL, and how maybe they can see it being improved and then again, higher level of stare some feedback. So, if nobody wants to get any feedback right now, completely understood, but look out for a type or if you can again give any feedback on the SCL that feedback will be shared with folks that are again, thinking about the SPL and thinking about reworking and refactoring of the SDL and AKASH a number of different ways. + +Andrew Mello: I have one little note on feedback. It's always been a side project to mine to just knock out in a weekend, but we're still missing a converter for dockerfile or docker compose to SDL. So, just a website, where you copy and paste your existing docker, compose, And boom STL, we can do it. it's just a converter. So that should probably be on someone's bucket list at some point. + +00:45:00 + +Tyler Wright: Absolutely, I know that there's a team in the Casa thought that's thinking about that, but a good call on Andrew. + +Andrew Mello: But yeah, probably the club guys. + +Tyler Wright: Does anyone else have any initial feedback that they want to share right now, again, don't feel obliged. I'm going to again, hand out a typeform that hopefully people can fill out and give feedback over the next couple of days, so that we can start to again, Think about take that feedback into consideration over the next week or so, with the Overclock Labs Core team as well as other folks, that may have Questions or want to see that feedback about the STL? + +Tyler Wright: again, I'll submit that inside, multiple discord channels, right? Next day or so. And then King people to see if they can again get feedback because I know people have message me and talk to me, privately about their thoughts. Is there anything else provider related to anybody? Wants to discuss before I let you all go. + +Tyler Wright: Cool, again, great discussions on updates From the Pretor team, great discussion on content, moderation, and a number of solutions that are being again, brought to providers on the Akash network. Thank you, Zach for volunteering to support the prayer team. I will continue to track that to make sure that you are + +Tyler Wright: And closer together on again, the management and Moderation API. And again, Andrew, thank you for the work that you've been doing. Please feel free to share that work again. I know that at the submission. Again, the next phase for both the Prey Tour tool and the Akash OS. Tool is testing whether it be the content, moderation, specifically or again just a tool in general before doing more wide share and just green lighting it as a tool that the community's news and over to the same thing with cloud most and all the various tools. So again we just want to do some internal testing amongst the Overclock Labs Core team and members of the insider and vanguards community that are, trusted technical folks, many of which are on this call. So, again, we'll continue to monitor the testing portion over the next week or so. And I know we have some volunteers that are willing to test both solutions and we'll just go from there. + +Tyler Wright: But again much appreciate everybody's time. This Wednesday. Feel free to reach out to me if you have any questions or concerns but again I'm going to be pushing over the next couple of weeks. So that'll be going to the next provider meeting next month. We got these solutions kind of well developed and potentially we can start sharing them at the steering committee meeting. not tomorrow but the following month and then again being a really good place to have a number of solutions at the Akash provider community can use So again, thank you all much. Appreciate everybody on this call. Please feel free to reach out to me if you need anything. There's again a validator upgrade happening in about 15 minutes or a little bit over 15 minutes. So let's talk online. But everyone have a great day. + +Tyler Wright: Thank you. Much appreciate everybody. All right, bye. + +Andrew Mello: So yeah. + +Rodrigo Rochin: See you. + +Zach Ross: Yes. + +Meeting ended after 00:49:02 👋 + +