diff --git a/sig-clients/README.md b/sig-clients/README.md index 22aa65e5..16b719c6 100644 --- a/sig-clients/README.md +++ b/sig-clients/README.md @@ -46,17 +46,13 @@ The goal of this SIG is to foster a community around each of these clients that | #11 | Monday, December 18th, 2023 08:30 AM PT (Pacific Time) |[Link](https://github.com/akash-network/community/blob/main/sig-clients/meetings/011-2023-12-18.md) |[Link](https://github.com/akash-network/community/blob/main/sig-clients/meetings/011-2023-12-18.md#transcript) |[Link](https://msr577epylf5qgo6aexm2xbblsreurfi2pgy5ed5ssdw6ztgkxwq.arweave.net/ZKPf_I_Cy9gZ3gEuzVwhXKJKRKjTzY6QfZSHb2ZmVe0) | #12 | Wednesday, January 17, 2024 10:30 AM PT (Pacific Time) | [Link](https://github.com/akash-network/community/blob/main/sig-clients/meetings/012-2024-01-17.md) | [Link](https://github.com/akash-network/community/blob/main/sig-clients/meetings/012-2024-01-17.md#transcript) |[Link](https://75iuscsrjioekyahvk6vad3taa5ikpweibvxs2rf4amjvzvj6gfa.arweave.net/_1FJClFKHEVgB6q9UA9zADqFPsRAa3lqJeAYmuap8Yo) | #13 | Thursday, February 22nd, 2024 09:30 AM PT (Pacific Time) |[Link](https://github.com/akash-network/community/blob/main/sig-clients/meetings/013-2024-02-22.md) |[Link](https://github.com/akash-network/community/blob/main/sig-clients/meetings/013-2024-02-22.md#transcript) |[Link](https://x5g5djwxzxixepgqsmxl6rupiipshcicdmq5bf2tjac35fjv5f6q.arweave.net/v03RptfN0XI80JMuv0aPQh8jiQIbIdCXU0gFvpU16X0) -| #14| March 27th, 2024 09:00 AM PT (Pacific Time) | | | +| #14| March 27th, 2024 09:00 AM PT (Pacific Time) |[Link](https://github.com/akash-network/community/blob/main/sig-clients/meetings/014-2024-03-27.md) |[Link](https://github.com/akash-network/community/blob/main/sig-clients/meetings/014-2024-03-27.md#transcript) | Coming Soon | #15| April 24th, 2024 09:00 AM PT (Pacific Time) | | | -| #16| May, 2024 10:30 AM PT (Pacific Time) | | | -| #17| June, 2024 10:30 AM PT (Pacific Time) | | | -| #18| July, 2024 10:30 AM PT (Pacific Time) | | | -| #19| August, 2024 10:30 AM PT (Pacific Time) | | | -| #20| September, 2024 10:30 AM PT (Pacific Time) | | | -| #21| October 2024 10:30 AM PT (Pacific Time) | | | -| #22| November, 2024 10:30 AM PT (Pacific Time) | | | -| #23| December, 2024 10:30 AM PT (Pacific Time) | | | -| #24| January, 2025 10:30 AM PT (Pacific Time) | | | +| #16| June 18th, 2024 10:30 AM PT (Pacific Time) | | | +| #17| August, 2024 10:30 AM PT (Pacific Time) | | | +| #18| October 2024 10:30 AM PT (Pacific Time) | | | +| #19| December, 2024 10:30 AM PT (Pacific Time) | | | +| #20| January, 2025 10:30 AM PT (Pacific Time) | | | ## Leadership diff --git a/sig-clients/meetings/014-2024-03-27.md b/sig-clients/meetings/014-2024-03-27.md new file mode 100644 index 00000000..5603cdd6 --- /dev/null +++ b/sig-clients/meetings/014-2024-03-27.md @@ -0,0 +1,531 @@ +# Akash Network - Clients Special Interest Group (SIG) - Meeting #14 + +## Agenda +- Updates from the Cloudmos team, represented by Maxime Beauchamp. +- Discussion on Terraform provider by Joao Luna +- Demos from teams collaborating with the Neta DAO team and Piyush Choudhary. +- Demo by Lukas Meilhammer on his current ideas. + +## Meeting Details +- Date: Wednesday, March 27, 2024 +- Time: 08:30 AM PT (Pacific Time) +- Recording: Coming Soon +- [Transcript](#transcript) + +## Participants +- Anil Murty +- B S +- Cheng Wang +- Envy | Neta DAO +- Fire:ies.ai Notetaker Nicolas +- Joao Luna +- Lukas Meilhammer +- Massv +- Maury De Jesus +- Maxime Beauchamp +- Piyush Choudhary +- Scott Carruthers +- Tyler Wright +- Zach Horn + +## Meetings notes +- **Tyler Wright** + - Welcomed participants to the March 2024 SIG (Special Interest Group) for clients. + - Noted the focus of the SIG on tools built on top of Akash to aid client-facing projects. + - Mentioned a busy agenda with updates from various teams building on Akash. + - Tyler Wright emphasized the importance of active participation and feedback during the demos. + - Noted the lack of traction on Piyush's LLM tool demo from the last meeting, suggesting community disinterest might indicate the project’s limited potential to grow the Akash Network. + - Encouraged attendees to provide feedback to help guide next steps for the presented ideas. + +- **Maxime Beauchamp (Cloudmos Team)** + - Provided a brief update on ongoing projects: + - Working on the GPU endpoint for the pricing page, coming online soon. + - Refactoring the ELBs (Elastic Load Balancers) in the Akash Console. + - Implementing provider monitoring for the new Provider Discovery endpoint, currently rolling out. + - Confirmed recent updates: + - New feature: Deployment details now show the GPU model being leased, improving GPU visibility. + +- **Q&A Session on Cloudmos Team Update** + - Tyler Wright invited questions or thoughts on Cloudmos updates from the participants before moving on with the agenda. + - Acknowledged the busy schedule of the Cloudmos team and their recent accomplishments. + - Encouraged participants to track Cloudmos updates via the project roadmap and engage with the team on Discord. + - Mentioned the regular participation of teams building on Akash. + - Noted a missing presenter but decided to continue with the agenda. + - Introduced Joao Luna to discuss making the Terraform provider more accessible. + +- **Joao Luna (Terraform Provider Update)** + - Provided background on the Terraform provider, developed about a year and a half ago. + - Explained that the provider is stable but underdeveloped, currently using the Akash binary. + - Initiated a discussion to open source components developed at Quasarch. + - Proposed moving the Terraform provider from his profile to the Akash Network organization to improve discoverability and trust. + - Planned to publish the provider to the official Terraform registry under the Akash Network. + - Detailed the next steps, including the open-sourcing of client libraries and improvements to the Terraform provider. + - Emphasized the need for secure generation and configuration of a GPG key by someone from Overclock Labs. + - Offered to assist with the process but stressed security concerns. + +- **Tyler Wright** + - Thanked Luna and opened the floor for questions. + - No questions were raised, so Tyler agreed to coordinate a meeting between Luna and the appropriate person from the core team to handle the GPG key setup. + - Confirmed the transfer process should take one to two days. + - Highlighted the importance of security to prevent malicious access. + - Proceeded with the agenda, noting two scheduled demos from present participants. + +### Piyush Choudhary (Neta DAO Project Demo) + - Introduced himself and the project aimed at reducing friction in onboarding developers to the Akash Network. + - Identified the pain points in the current process of deploying GitHub repositories to Akash. + - Proposed a streamlined process, similar to Vercel, to deploy applications from GitHub to Akash without manual Docker and SDL file creation. + - Showcased a POC (proof of concept) that automates deployment scripts. + - Emphasized that the solution will support multiple languages and frameworks, not limited to JavaScript. + - Demonstrated the automated deployment script detecting changes in the repository and redeploying without manual intervention. + - Mentioned ongoing discussions and proposals for this concept. + +#### Q&A Session on Piyush Demo +- Anil Commented on private repo support being code complete and asked about Piyush's webhooks depending on container image tags. +- Joao Luna asked if the deployed image actively monitors the repository for changes. +- Joao Luna Piyush confirmed that the image checks for updates every five minutes, with plans to use GitHub APIs for real-time updates. +- Scott Sought clarification on the goal of Piyush's solution: +- Whether it eliminates the need to customize an SDL. +- Whether users need to be knowledgeable about SDL as part of the deployment process. +- Scott confirmed understanding of the Docker file automation. +- Scott inquired if the solution also removes the need for manual SDL customization. +- Scott Suggested a feature to auto-populate SDL with appropriate resource settings for AI models, reducing cognitive load on users. +##### Piyush Choudhary: +- Clarified that the solution does not depend on version tags in the SDL. +- Explained that the environment builds are managed by background scripts that handle framework detection and deployment commands. +- Explained the solution's goal to simplify deploying applications from GitHub to Akash: +- Eliminates the need to create Docker files manually. +- The solution manages the deployment process and updates. +- A GUI will allow users to log in with GitHub, select repositories, and deploy without creating an SDL manually. +- Users can see a preview of the SDL, but it is auto-generated based on the repository and environment variables. + +#### Demo by Lukas Meilhammer +- Noted that there were 22 minutes left and handed the floor to Lukas and Maury. +- Mentioned that Lukas's team had been active on Discord, addressing technical challenges while deploying on Akash. +- Lukas introduced their project, a data and knowledge marketplace allowing users to monetize or control access to their data. +- Lukas highlighted key features such as private or public databases, token-based access, and infrastructure for reward distribution. +- Explained the integration with decentralized cloud services like Store J. +- Presented a demo of the platform, emphasizing the user interface for database management and access control. + +**Integration with Akash** +##### Lukas Meilhammer: +- Tyler requested clarification on how Akash is being utilized in their deployment, focusing on practical implementation and related issues. +- Described their current deployment architecture, with different providers handling different components: + - Frontend on one provider. + - S3 storage gateway on another. + - Backend and database on a third provider. +- Mentioned challenges with Docker images and potential solutions. +- Expressed interest in integrating solutions discussed earlier in the meeting to streamline their deployment process. +- Invited feedback from the community and highlighted the need for deployment best practices and clearer roadmaps to reduce friction and redundancy in resolving issues. +#### Maury De Jesus: +- Added details about the current deployment status on Akash: + - Entire platform is running on Akash. + - Database is stored on Store J. + - Frontend, backend, and S3 gateway are integrated with the Akash network. +- Opened the floor for questions. + +**Questions and Discussions on Lukas Demo** +- Anil Murty asked about specific issues faced during deployment and how they were addressed. +- Scott Inquired about the user experience for managing databases on the platform and how Akash improves this process. +- Piyush Choudhary asked about backing up deployments that go down by the providers. +##### Lukas Meilhammer: +- Elaborated on deployment challenges: + - Inconsistent performance across different providers. + - Difficulties with Docker image management and redeployment. +- Acknowledged that the solutions discussed by Piyush might help address these issues. +- Emphasized the need for an API to streamline deployment and update processes. +- Explained that Akash provides a decentralized, scalable infrastructure for their platform, enhancing data security and availability. +- Mentioned that using Akash eliminates reliance on a single cloud provider, thus reducing risks and improving uptime +- Acknowledged the need for regular backups. +- Mentioned they have POC clients and partners looking to launch or update data. +- Expressed willingness to schedule calls for further discussion and technical collaboration. +**Next Steps:** +- Lukas and Maury's team to explore integrating solutions presented by Piyush to enhance their deployment process. +- Community members to provide feedback and suggestions for best practices in deploying on Akash. +- Continued discussions on Discord to address challenges and share updates. +- Tyler highlighted the availability of a private Discord channel for technical discussions with Lukas and the team. + +### Backup and Data Management Discussion +#### Piyush Choudhary: +Asked about solutions for backing up data and whether Lukas' team has implemented any backup routines. +#### Lukas Meilhammer: +- Mentioned ongoing efforts to identify errors and the absence of a backup routine earlier. +- Invited collaboration and technical discussions to expedite solutions. +#### Anil Murty: +- Clarified that the front end and backend are running on Akash, while data is stored on Storj. +- Emphasized the need for regular data backups since persistent storage on Akash does not automatically backup data. +- Suggested implementing a backup solution on their side. +- Emphasized the importance of backing up data from persistent storage on Akash. +- Suggested regular backups to Storj or another solution. +- Offered to connect Lukas with reliable providers on Akash for further discussion. + +## Action Items +- Tyler Wright to set up a conversation between Joao Luna and the appropriate person from the core team for secure GPG key handling. +- Follow-up discussions between Piyush Choudhary and Anil Murty on private repository deployments. +- Lukas Meilhammer and Maury De Jesus to potentially collaborate with Piyush Choudhary on deployment automation and best practices. + +# **Transcript** + +_This editable transcript was computer generated and might contain errors. People can also change the text after it was created._ + +Tyler Wright: A Wright welcome everybody to the March 2020 for addition of big clients. This special interest group for clients is focus on all things related to those that are building clients facing tools on top of Akash. I know there are a number of projects that are being built on top of clients to make it easy for people to use a cash and so again, That allows people added to their workflow. so much appreciated everyone's participation and constant participation during this sig. + +Tyler Wright: There is a pretty good size agenda happening for today. So just out of the last meeting usually during these Sig clients meetings. We get updates from various teams that are currently building on top of Akash. We can update from the cloud most team take Max is here representing that team on all the stuff that they've been doing. Following that update. I know that Luna has been working on terraform provider and put a discussion number 502 which I'll drop in the chat. + +Tyler Wright: To make that more accessibly open source, and so we'll talk a little bit about that. Following that conversation there are a couple of teams that I've asked to join this call to give short demos on ideas that they've been thinking about inclusive Concepts that they've been working on one. Those teams is part of the Dao team working together with piyush and so they have an idea to go again give a high level demo on and talk about and then Lukas who's also the call has been working on some ideas and so he will give a demo as well. So again out of the last + +Tyler Wright: Said clients' meeting. There was a demo given by a Pearson is his team related to an llm tool. There was not much traction on that discussion afterward. So we're hoping that people on this call can participate actively it does not much attraction on idea. I had to imagine that means that it's something that should not be pushed forward because the community doesn't see much need for its Late as it stands to grow the Akash Network. So again the folks that are participating and sharing some of those ideas later today in the agenda are certainly looking for feedback to take necessary next steps. So just want to throw that out there. + +Tyler Wright: So kicking off the agenda. I'll hand it over to Max who will give us some updates from the cloud most team so max feel free to take it away. + +Maxime Beauchamp: It's gonna be a brief update. We are working on a lot of stuff currently. But yeah, we'll be working on the GPU endpoint for the pricing page which will be coming online soon. We are. Continuing our work to refactor the elbows deeply into the cash console. Which will also come in the coming weeks and + +Maxime Beauchamp: a bunch of stuff for provider monitoring for the new provider Discovery endpoint, which is a rolling out as we speak. So yeah. That's pretty much it. + +Tyler Wright: Over the last month as well Max. I know they caught up in the thick of all I think they're working on right now. But over the last month you all have had a couple of new items in terms of how people could view gpus on the clubhouse web app that correct. + +Maxime Beauchamp: Yeah, we released on the deployment detail you can see which GPU model you are currently leasing, which was not possible. previously So, yeah, it's a big Improvement for GPU visibility. + +Maxime Beauchamp: Thanks for the reminder. + +Tyler Wright: + +Maxime Beauchamp: Probably something else that I forgot about just the rush right now. + +00:05:00 + +Tyler Wright: yeah, no, no worries. Again, a lot of the work of the clubhouse team does is on the project road map for the product and engineering team so you can kind of track a number of those releases. The clubhouse team is also quite accessible at Discord a number of channels. But if you have anything related to something you want to build on top of Cloud most or questions specific to Cloud most in general the ecosystem Cloud most channels a great place where again anything that's like an issue currently with the cloud most tool you could drop in there and then an issue will be created inside the + +Tyler Wright: Was repo and GitHub and the Really responsive about solving those issues as well as ideas or questions about the tool anything else. You can certainly ask in that channel amongst many other channels. Does anyone have any thoughts or questions about Cloud most or anything? They want to ask Max about related to Cloud most updates at this point before we continue on with the agenda. + +Tyler Wright: All Again, there are a number of other teams that are currently building on top of Akash every month I ask those teams to join and happy to move this meeting around to accommodate folks schedule. There was somebody that was scheduled to come to talk about status of their project but has not attended quite yet. So we'll continue on with the agenda. as mentioned at the top Luna who's on the call here had previously been working on terraform provider for a while and I think Is talking about making it more accessible for other people to contribute to the growth Luna. I'll hand it over to you and then we can talk about any next steps if necessary to make it available. + +Joao Luna: Yeah, So for those that don't know the perform provider it was about like it and year and a half ago since then it has been. quite stable and one of the reasons and it's kind of ironic one of the reasons that's that stable is because it's not that Developed so at the time there were really no clients. So we use The cash binary in the system and as long as the binary works and the people in the users keep it updated the provider will keep running with no problem. This has some challenges in the maintainability as well. So + +Joao Luna: the discussion that I started is the first step in a series of efforts that will be doing to open source some of the components that we developed at quasarc that make our platform that we're developing possible. But there are also benefit the Frog community and we think makes sense open sourcing the terraform provider is already open source, but it's under my own profile. The idea of these first step is to move it to the + +Joao Luna: Cache Network organization and then as part of the process moving to That official terraform registry, but under the cache Network organization instead of my username. So it will improve discoverability and Trust as users do find it weird to download an official. Let's say package from terraform. That's not under the official organization. So this is a step towards the debt. And then the next steps will be the kai libraries that's been due for a long time. It's just that we haven't had that Focus yet to open source it's a coupled to our project and we've been decoupling and might soon be ready to open source, and then as part of these efforts, there'll be an effort to further effects that are from provider using that client library that we will release + +Joao Luna: so the code base gets overall more structured and better. And these are the three next steps on that list being the first one that donation of the terraform provider and it will consist basically of moving the terraform provider to the cash organization and then some technical stuff to actually publish it to the terraform registry through the GitHub actions. So what we'll need will be to generate a gpg key. I assume this should be done. Securely by someone from the overclock team that will be able to generate this key. I can walk you through the process pretty easy. But yeah, then this will have to be configured in the repository this gpg key and then GitHub actions will automatically + +00:10:00 + +Joao Luna: publish the artifacts into the registry. But yeah this process I'm available to do it, but I think I should have someone from the overclock really lab team. to actually Configure the gpg key securely, so I don't have access and things can be published more securely. and I think this is the overview of that discussion and the further ones that will arrive in the near future + +Tyler Wright: Perfect. Thank you Luna for that context. + +Joao Luna: any questions + +Tyler Wright: Yeah. Does anyone have any specific questions or concerns about next steps that Luna's Outland? + +Tyler Wright: Cool. If not, then I'll set up some conversations and private conversation between Luna and the appropriate party from the core team to handle that. + +Joao Luna: This process shouldn't take more than one day. So just for context what no more than one two days and… + +Tyler Wright: Okay. + +Joao Luna: we can have this transitioned. It's just that it should be done securely because whoever has that private key can publish an artifact and we don't want that key to fall into malicious ends and publish something malicious in name of the Akash Network. + +Tyler Wright: No, that makes Much appreciate that. So I have an action item that again. I'll set up a conversation with the appropriate person to get that handled between meetings. But thank you very much. And if anyone has any questions comments again, it'll be announcement made. What's the transfer is done for people to access and contribute. + +Tyler Wright: Cool, if there's no questions there then we'll continue on with the agenda. Cool. I know that there again two scheduled demos. I see from folks that are currently here. + +Tyler Wright: I will hand it over to P issues been working with the net of Dao team on a project and PSU can share your screen. If you need to I'm hoping that these demos can be as discussed less than five minutes, talk about what you're trying to solve at the highest level again If you have one how this will improve the accost Network and then we can get into more of the nitty-gritty be at some questions because there's many folks on this. Call that our combination of Technical and non-technical. So again, if we can keep it high level and fairly focus with these demos, then we can get into the nitty gritty and questions and elsewhere, but I'll hand over to you piyush. + +Piyush Choudhary: Yeah. Yeah, thanks for the introduction Ty So those who don't know me so I'm pierced from digital. So we have both for official question networks, exactly so we have been exploring how we can reduce the fictions in onboarding the developers on the So one of the major pinpoint in Akash network is to deploy the GitHub River State I click on GitHub onto Akash. So there is a normal process of getting Dr. Files then creating an HDL then going over to Cloud models and depending it over to the cost and also to update the deployment of nakash first even to get abusive getting to work close the image gets belt and then the user have to go to the console again and updated plan and basically so there is a lot of predictions in between the process and applying applications. when we compare this with the complete first ourselves and + +Piyush Choudhary: London Network, so they have a clear pipeline where they can log in with GitHub select a repositories and they have kind of template or you can see the basic steps which takes care of all the requirement process in case of versatile. They use AWS and the sales and they just run all the scripts automatically without worrying about creating the docker files or basically HDL Etc. So this is what the idea is. Basically we are trying to streamline the process of deploying and applications from GitHub without cash without worrying about the dog creating Docker images or writing Industries. So people have graphical user interface on the cloud moves which will have the single features and then we can have the multiple Spirits for example, in case of versus limited to deploying JavaScript libraries, but with our concept as a POC we can do this for python Library. She can do this for JavaScript anything. We just have to write a proper that's behind the scenes that take years that it takes the Frameworks and basically + +Piyush Choudhary: States there of the deployment commands in the loss. So these quick POC. I would like to Showcase and also for more information says can go through the discussions or drop discussion that I have pushed. and I have a quick POC. This is a quick PRC to get around the idea slowly drive and more updates will be sure to the proposals. So I'm going to like them okay minutes, so Let's say so right now we are going to the steel and in the docker image. We have kind of created a scripts that takes place of everything. So let's say if I want to deploy this so this will take some time as it's big. Let's deploy this + +00:15:00 + +Piyush Choudhary: So some instance there are some errors in this as it's a POC but as we work around so it will be more beer. So as you can see this is the documents that we are created in which it takes care of all the scripting detection commands that runs in the locks and whenever also in case of course, right so we talk about the computer whenever we make the changes to the data presentation. I always thought comes in the GitHub that says the deployment has been triggered on Versa and it automatically pulls the changes and deployed onto the worst. But in case of the car it's an animal process. So this will also solve this. So as soon as the genius something push the greater purpose Jesus origin to I learned that the deployment has been trigger on the cloud most or it's Akash and once successful we have a techno. This is all going to be done to get out and if this + +Piyush Choudhary: So feedback Also suggestions are more than welcome. And I hope this will bring solve the major friction that the current system have maybe Act has an additional layer which will solve the deployments of repository from you don't want to a cost and it will also allow to deploy private repositories as well. Likely from GitHub to cost to get So this is a foundational structural that we have made and the drop proposal is also being pushed. for discussion + +Piyush Choudhary: So till this can deploys anyone have any questions. + +Tyler Wright: I know there's a couple of questions in the chat about private repo. I know that that's something that is actually code complete from the core team and is going to live as it's a core feature to the Akash network over the next week or so. But anyway, I saw your hand up Anil. Do you have a specific question or a comment that + +Anil Murty: Yeah, comment was same as what he just said Tyler. So the core team has been working on supporting Data Bridges fees so we should have a discussion about that and see which solution we ultimately take. so that's number one and then the second does Your Web book essentially depend on being able to utilize a tag on the container image. So I noticed in the STL you have a version tag new depend on that + +Piyush Choudhary: so basically the tag it's not dependent on this. So if you check the STL so we have the environment belt. So whatever you say that goes into goes in the environment third and the image that is being assigned as still is nothing but the background scripts that takes care of the detection and also runs the proper commands of detecting the proper Frameworks similar to what versus + +Anil Murty: Got it here. The reason is mentioning that is because we generally don't recommend people use version tags in the sdl… + +Piyush Choudhary: we have + +Anil Murty: because of the way that we have configured kubernetes. + +Piyush Choudhary: so I think there is probably + +Tyler Wright: That's Luna. Did you ever your hand up? + +Joao Luna: Yeah, it's just a quick question. Perhaps I didn't catch that. Super Mario understood what you do is deploying an image, right and that image will be actively looking for it link repository that you have there. and basically Have your application running right that repository running? is Sorry. + +Piyush Choudhary: So you could check that. Yeah, you can continue. + +Joao Luna: Is that image actively looking for that repo? + +Piyush Choudhary: Yep. + +Joao Luna: So when I make changes? Okay, cool. + +Piyush Choudhary: So we have a timeline set right now it's five minutes. So every five minutes check all the changes but as we implement the GitHub API six will be able to detect as soon as The Pusher are being made to the represents. + +00:20:00 + +Piyush Choudhary: As well the demo. I think it's failing for some reason. I think it's building we don't need. As you can see the script. + +Tyler Wright: Anil I know you had it + +Tyler Wright: I'll go ahead to Sorry, I know Anil has his hands up and I'm not really sure if you answered his previous question about versioning, but go ahead and then I'll hand over to Anil again. + +Piyush Choudhary: So as you can see the script was able to detect that it was very clone the reposites and on the appropriate commands. And as far as we explode computer computers also to the similarities. Yeah, but in short it will not be just about The Versatile thing it will be about any type of things like we can also get a pipeline to deploy button applications. for example, right now aawpers uses python libraries a lot, right so we can have a direct a pipelines which can deploy the python based. applications directly from get one to across without worrying about meeting the local fire or SQL. It's Okay, so this is being deployed as you can see I can just go to. + +Piyush Choudhary: Okay, so this is being declared without creating the sdo 5 so as small questions. Right now the image is just for the POC. So this will Is behind this is nothing but a static or wait a script which takes care of all the deployment by clients. It's kind of automated feature. You can say additional layer which acts as a manager which takes care of all deployment pipelines. And as soon as the dates are getting pushed to the reposites, it keeps on if it keeps a watch and then really deploys the thing in the locks. So yeah. + +Tyler Wright: I see anil's hand up and then Lukas following. So no you go. + +Anil Murty: Hey, Yeah, so a couple of questions one is So then would each clients essentially be deploying this along with their images and pointing it to their repositories or how would so that we want clients per provider. So let's say there's 10 workloads running on providers. There were 10 versions of this running or how this work. + +Piyush Choudhary: So basically, of course the 10 version of this will be running. So right now it says deal so instead of STL we will have a GUI. So in the GI, we will have to options login with GitHub which through Earth we will load all the requests into the platform and say it deploy and The children will not be visible to the users standpoint, but behind the scene as Akash Independence completely on their steel what we are doing through GUI is we are just creating the STL. Under the hood with the repositories URLs and whatever you said, so let's say 10 are trying to use the stool. So basically those 10 users will be deploying the 10 SQL files. So yeah, I think that runs your question. + +Anil Murty: Okay, got it. And then you talked about in supporting private Registries is a privateer GitHub repositories as well. not Registries. Where are you passing the auth information to? + +Piyush Choudhary: So basically now all information should be passed into the logs. It's a + +Anil Murty: How that works. So, ultimately how does your specific app get the auth information + +Piyush Choudhary: So basically when you create a login with get a brighter so that through GitHub apis we can all the represents into the platform. + +Anil Murty: always + +Piyush Choudhary: Yeah, so yeah, yeah. + +Piyush Choudhary: Yeah. + +Piyush Choudhary: Yeah, so we got the key as the login with get up through Earth. We get access to the repositories as there are permissions that we can control through GitHub apis. And as this is the best basically is a manager of all the things when it starts running the commands in here so it takes care of the old Logan it can deploy the private depository as well in that way. So we have to research a bit around this as we have just started working on it. So it was just a POC. So as the foundation gets more clear. We will have more clarity on how the things to look. + +Tyler Wright: Lukas, I see your hand up you have a question. + +lukas meilhammer: Yeah. Hi This sounds like it's touching on a problem that we also had there was a unexpected downtime and in some cases. We didn't know why our application was down. So I mean we have the back end in the database in one rep in one provider and then store J Gateway and front and on another provider, maybe generally Our architecture could be reviewed but we also had the problem that the provider had to then actually do some node maintenance and one time in our developer Maury in the call. First of all, we will have to review your project to you. And then we might also want to get back to you or… + +00:25:00 + +Piyush Choudhary: Yeah. + +lukas meilhammer: jump on the call if that's needed. But + +lukas meilhammer: Lovely to hear that private repositories. I mean coming into the priority a little bit of prioritize a little bit more and… + +Piyush Choudhary: simply + +lukas meilhammer: maybe we wanted to instead of deploying the application via cloudmoss. We wanted to just do it via CLI as So it's the console that you are talking about I bet yeah. + +Piyush Choudhary: so I think of course. Yeah, you can continue that I would. + +lukas meilhammer: Sorry. + +lukas meilhammer: No, please and Maury has actually got more insights on that and maybe it's something for another call for more. Let's say special interest group, but let's please get together on that because we were thinking of let's say collaborating on these kind of things. because we're going Into a phase… + +Piyush Choudhary: if shows + +lukas meilhammer: where we want to do a private or even soon a public data of our platform, which is a data management and data curation and moderation platform. That can be used for many things. So one proposal was actually to do a best practice knowledge base on our platform for Akash users in general. + +lukas meilhammer: Then maybe to have moderators actually changing things in the knowledge base. which could be really a documentation and additional one. And generally maybe I would like to showcase the platform. But let's first finish this topic off and I wanted to just say that we're working on a solution. We have a script to redeploy on the app if it's down and we want to automate the process and figure out where exactly a problem. The occurring from to actually maybe just issue and automatically restart containers and all that. So Maury do you have any thing to add right now related to this discussion here? + +Tyler Wright: Before you all go because I want to give you all some time real fast to talk. + +lukas meilhammer: Yeah. + +Tyler Wright: It's either introduce yourselves a little bit and talk about what you're looking to build a pH. I see your hand up. But real quick. Does anyone else have any other questions or comments at this time about? What piyush just presented? + +Tyler Wright: Go ahead Scott. + +Scott Carruthers: Appears to be you bring up the STL again. + +Piyush Choudhary: If so, yeah, give me two minutes. + +Piyush Choudhary: says nothing visible + +Scott Carruthers: Yes, and then can you so let me make sure that I understand the intense. I understand the removal of the need to create a container image. is a solution also trying to negate the need to customize an sdl. Or so I thought I heard that maybe in the beginning in the presentation, but maybe I'm up. So in order for a user to be able to use this service. Will they need to be? Knowledgeable about a Nicolas sdl as a part of this process to negate the need for a user to have to manually manipulate a sdl. + +Piyush Choudhary: So yeah, so I will take on that. So basically what we have done here is in short the image is the steal anything but a step that is acting as a manager deploy the bit of applications from get up to Akash and also running the commands in locks right after closing the equal States now the thing is like + +00:30:00 + +Piyush Choudhary: Users will always be able to see so we have a te So if I go to the templates not downloads. So as you're going right in case of azure, but we have the first to basically manipulate their skills. GUI, and then we can preview the stove right and single case. We will also have the thing right now here the users do they do this with their own images instead of the images. So the documentation it will not be there with our solution because we see the docker images something that is automatically the things so here we will have the GitHub reposition and if they login with GitHub, then they will have the access to basically whichever it was Jesus. They want to learn to a car and then all these things will be like again, I think the memory and then also they can add things and then what will happen is in the preview for the image will be ours that is + +Piyush Choudhary: you solve which is nothing but you can say a script that is taking care of all the things and then We have the GitHub Registries the environment variables related to the bit of the pastries similar to if you go to the person there is a fair to add environment So yeah, this is something that is eliminating the need for the users if they want to deploy the applications directly from GitHub to Akash. Our solution is eliminating the need to basically create a Docker file and then + +Piyush Choudhary: to go through the process of getting your local file and then deploy you don't do a cost by creating SQL file. And for example, whenever we make the updates to the GitHub Webster's creates a new image who either work close or by manual then to update that employment on We have to come here and basically change the image. Tag and then he deploy to update the deployment. But in our case our script is speaking of what's on the reversies as soon as the push has changes it automatically schools the changes to logs and then they deploys the thing about the latest update. + +lukas meilhammer: So it shortens the seconds and eases the updating General and Autumn is basically actually giving us the option to great. + +Scott Carruthers: Yes, I understood the piece of taking the authoring of the dockerfile. So I understood that when I talked about the sdl What I was curious about is this also taking the needs. I understand you could arrive at a club must screen where you can manipulate the sdl and manipulate the amount of resources that you're throwing at a deployment. I was wondering if this was trying to solve taking that cognitive load off of the user as well. So if I was going to deploy a AI model onto a caution, I was using this service that it would automopulate the sdl with appropriate GPU models and appropriate CPU and RAM and storage. Is that part of this or is it just taking off the load of creating the docker image? But the user would still be responsible for authoring the sdl appropriately. + +Piyush Choudhary: So basically we can consider yourself around this I think this is a possible to this daily prefill the resources that is required to plan applications from getting one to Akash. So we have to research a little bit on this right now. + +Scott Carruthers: Okay. Yeah. + +Piyush Choudhary: It was not a part of this but we will do some research together on this. Yeah. + +Scott Carruthers: Yeah, that makes sense. Yeah, I mean it's very powerful just reducing the cognitive load of not having to create a Docker file. So there's a myths value in that in and of itself. I was just wondering if I was also trying to reduce the cognitive load of being familiar with the sdl but understand that's not part of this current version but maybe something that you would Implement and… + +Piyush Choudhary: Yeah. + +Scott Carruthers: Future versions are research there. Okay. + +Tyler Wright: All right, it seems like again it's been great discussion around this. I know next steps are not a doubt team in piyush are gonna put a discussion around this I would urge that people could get involved in that discussion and give feedback ask questions poke holes so that again those teams can answering those cars and make everybody feel comfortable but before taking any other steps beyond that looks like a discussion is already live. Thank you very much. So again, + +Tyler Wright: I want to make sure and I'll continue to share this but hopefully the net a Dao team and piyush can continue to share this across Discord get feedback from members of the community in general Chad and all channels and just continue a lively discussion around this. That would be very impactful. But again much appreciate the time that you took today the work that went into this and thank you everyone for those questions. + +00:35:00 + +Tyler Wright: All I know we only have about 22 minutes left. I wanted to give a Lukas and Maury some time to talk. This is a group and talking to a bit Discord and Elsewhere for a little while and they're a very technical group. That's again building their own Solutions based in Europe. I want to In the platform to talk about some other they've already been deploying on a caution. It's found some issues and again and working to solve those issues. So I want to hand over the floor to Lukas to talk at a high level, about themselves what they're building the advantages to the Akash Network and then we'll open up to questions and get into the weeds. + +lukas meilhammer: Yeah, hi everyone and here in the Juventus touring. No trickle one of our code developers Maury our city hall can join how today but We are basically building and I'm just going to bring it up probably almost directly. We're basically building a data and knowledge Marketplace. We're trying to let the user decide how they want to monetize their data or how they want to give access to it. So whether it may be token holding having an Nifty and basically paying a subscription doesn't really matter too much for us because we are basically building a little bit of instruction infrastructure around this. + +lukas meilhammer: Taking contracts to issue rewards at some point all that. So it's a no-code database built on a decentralized cloud storing the data on store J. + +Piyush Choudhary: let's + +lukas meilhammer: It's been quite a bit of a built so far, but we're slowly moving ahead and We are in the process of actually launching a public or private data or I mean databases can basically be entirely private protected or public so I will just show the platform quickly + +lukas meilhammer: I hope to just gain some attention and get some people on board that want to build with our platform. So. I would say. I'll just walk you through the different top level panels here. You have a profile you can browse users in the next version profile can also be private so that you don't see the actual wallet hash that someone logged into then you have incoming requests for the database so that you can basically add someone's use a hash and give him access that usages something for later. + +lukas meilhammer: And then you have incoming outgoing and you have changed requests because basically in each database, let's just open one up quickly. This is a protected database which basically means everyone can see this page. Everyone can see. The criteria the description everything of the database and there should actually be the data schema down below here. So the data schema is public the metrics on how the platform is used here. We see an error already which sometimes we don't know exactly where it's coming from. And then basically it's more related to its. Hope that opens at least now. okay so I can basically make inputs in this field private with a right click + +lukas meilhammer: And then it's going to appear gray. And for the user that is not the owner. It will just be empty and empty field. you have all the filters that you usually have. You can set up a form you can change the form. So it's basically like an air table on a decentralized cloud with giving everyone the options to monetize the data day as they wish. Tyler you have a question already. + +Tyler Wright: Yeah, I appreciate you showing us this I just want to make sure because again many of these folks on this call are focus on Akash. + +lukas meilhammer: Yeah. + +Tyler Wright: I just want to see where the implementation of their cash network is and I remember you talking about redeploy tools as you were leveraging a car. So I just wanted to make sure we got to that part because I'm sure there'll be some questions and yeah. + +lukas meilhammer: Yeah. + +lukas meilhammer: So I'm not a CTO but let me try to summarize this a little bit. So as I said, we have the deployment a little bit split between providers front banners on one S3 storage Gateway is on the other on another one and the database and the back and is on another provider so correct me if I'm wrong here Maury. + +00:40:00 + +lukas meilhammer: But generally we were facing some problems also with the docker images and all that and we just have to maybe review the solution that was presented here today to figure out whether this already solves our need but going into a public beta in we might want to have that iron out and we're happy to wait a little bit. But generally I'm here to get some feedback ask people on whether they want to build something with us and maybe generally also try to let's say summarize these calls sometimes a little bit because + +lukas meilhammer: there is a little bit of friction and maybe questions that are answered multiple times in the Discord. So thinking about a deployment best practices or maybe even a more let's say open roadmap on what's about to come To just actually not worry about stuff that you don't have to be worried about because there's an update in one week. For example, that was one approach where we thought we can use the platform as it is, but that would of course have to be discussed in terms of what else would we need an API that pulls data from other repositories where you document stuff so thank you for now born in first intro call. This was really helpful and all basically. + +lukas meilhammer: so what's running on Akash, everything is running on a cash and the database is on storage a doll the backend the front end and S3 Gateway for + +lukas meilhammer: For the storage integration that the S3 store there was this code. Yes, please go ahead with questions. + +Piyush Choudhary: Yeah, so I have a question. + +Tyler Wright: Go ahead. Peace. + +Piyush Choudhary: So it's an initiative talk about backing up the deployments that caused town by the providers. So I would like to ask where is that thing going on? And also you talk about you lost all the data and the deployment goes down from one of the providers. So are you obviously teaching the solutions that backs of the data or is it something else? + +lukas meilhammer: My meeting already came out for this but there was no backup routine. we have definitely in the process of identifying the errors somehow or in any way. We are also happy to share the insights. And if there's really something and that you already working on and we were of course to know a little bit ahead of time then when it's uploaded or maybe take some part into the actual definition or scripting or work and then this could all move ahead a little bit quicker and I'm generally a very big believer in decentralization. So happy for any collaboration that pops out of this calls here. Anil + +Piyush Choudhary: Yeah, sure Yeah, so I will bring you upon Discord to see if you can call them collaborate on some things type you're working on and other hand over 200. + +lukas meilhammer: Perfect. No, please let's schedule a call latest next week or so Maury. Do you have anything else to say and before Anil? Maybe + +Maury De Jesus: For now and it's okay. Yes. + +lukas meilhammer: Anil + +Anil Murty: Yeah, so this is cool. And I didn't realize I only heard storage in S3. So I wasn't sure if something was running on So it seems like the front end in the backend are running on Akash and all the data is actually stored on store Jesus. Is that right? Okay. + +lukas meilhammer: And it's only data. + +Maury De Jesus: Yes. + +lukas meilhammer: So there's not a lot of data traffic and all that. I'll just basically forward you because someone was asking me for the provider. I don't think that there's any Critical information just post the whole thing into the chat. Really you quickly know just + +Anil Murty: Got it aakash proba Okay, so I was mostly curious about which providers you're using because I think so. There's a couple of things I don't know if you're familiar, but essentially Even if you're using persistent storage on a cash today meaning stories that person essentially still have to make sure that you back up your data somewhere else. + +lukas meilhammer: Yeah. + +Anil Murty: So whether it's true or just something and there is no really backup solution today that's available and that's potentially something for Either the core team or somebody in the community to look at I've been on our list of things but just haven't gotten to it but it's a very common request that most customers have the recommendation because you implement something on your side backup your data regularly because when the lease closes you lose it. so that's basically so as long as you're backing up the data to storage it should be okay and then + +00:45:00 + +Anil Murty: At least from what I understand today, you don't have active users on this product,… + +lukas meilhammer: Yeah. Yes,… + +Anil Murty: right? It's still in. + +lukas meilhammer: no, no, our puc clients Partners actually looking to launching or updating some data. I don't necessarily see them because they are not shared with me or… + +Anil Murty: Sorry, yeah. + +lukas meilhammer: whatever, but basically if you have a database you can keep it private. Then you make it protected so that people see the data schema you get maybe initial requests for Access and maybe some more insights. + +Piyush Choudhary: just + +lukas meilhammer: Yeah, and then basically at some point maybe you want to put it public because it's got some marketing value what not. So generally, we really happy to also please also add me on LinkedIn. Maybe we can just schedule a call also a deaf call or Tech call for anyone my class interested in,… + +Anil Murty: Yeah, sure. + +lukas meilhammer: Idaho. all that + +Anil Murty: It didn't actually just to preserve time on this call. I'll just have you LinkedIn and we can have a conversation about this and see we can help it in terms of,… + +lukas meilhammer: Perfect. + +Anil Murty: giving you better providers or looking at connecting with providers that have been on the network for a long time and then we're gonna have a user confidence. So you want to talk about your solution there to the greatest. + +lukas meilhammer: If you don't mind, I found a few of you guys on LinkedIn already. I will just maybe make a connect request. + +Maury De Jesus: Try again to connecting in French changing Lukas. + +lukas meilhammer: … + +Tyler Wright: and we also private Channel. + +lukas meilhammer: because the net that + +Maury De Jesus: Is refresh the fish. + +Tyler Wright: We also have a private Channel with Lukas and the team on Discord for admins that we can also talk in so much appreciate. + +lukas meilhammer: Yeah, what we do invite you guys there if it's for technical technically related stuff refreshed refresh this page. Yes. Now it's not so super important. I think everyone gets the platform and I mean eventually on the long run we would really like to rather make it a easy interface just like Google you search instead of websites, you get databases stuff like that. and then maybe you don't have access you still need to request or maybe you perform a quest or whatever how you want to monetize your data and I think with Jesus we will find a way to monetize and I'm really happy to have people meeting their own platform. I'm not really keen on offering white labeled versions of it at the moment just so that our brand and our project just gets more known out there but happy to we're working on the API actually pulling data in so that eventually you could have more like cryptocurrency. + +lukas meilhammer: Analysis dashboards with the market cap updating and all that stuff. Yeah, that's a very common request at the moment because we're in a bull market. So yeah. Yeah, I see that. Yeah. Yeah all good guys. Thank you for the flowers. + +Tyler Wright: this all right. + +lukas meilhammer: up just by the way, so + +Tyler Wright: excellent + +lukas meilhammer: Really? Excellent. I thank you for the time already and happy to invite you all to a group or start some I tell telegram Channel. Just shoot me a message on LinkedIn. I've been sending some requests. I'll just now. + +Tyler Wright: Excellent. All… + +lukas meilhammer: through joao zip + +Tyler Wright: I know we only have a couple of minutes left much appreciate again the demos from piyush and the net a Dao team as well as from Lucas Maury and the folks over at ipal again. We try to take these Sid clients meetings and I will adjust the schedule to make sure that more beings like this are available. But this is a great opportunity for teams to connect with one another for folks to talk about ideas that they're building and still they can get feedback from members of the community and members of the core team as early as possible so that we can continue to push. + +Tyler Wright: Or it's again demand driving usage on the network. Those being the main Theses that again everybody from the community hearing committee and elsewhere is agreed on how we continue to push forward the Akash Network, especially through clients. Cool. Is there any other questions comments or any other discussion points that anybody wants to bring up with just a couple minutes we have left. + +Tyler Wright: If not, then again, I'm sure they'll be following conversations with the ipal team. I thank you for joining. If you have any other thoughts, please feel free to drop them in the Sig clients' Channel, again or one of the other channels but said clients is a great place to put things so that everybody from the community can interact and give feedback because again there's a number of great teams individuals that have been around for many many years almond are very technical and can provide a lot of support So as much can be done in public would be great. But again, we'll continue to have conversations again. Thank you to Pearson that a doubt team if anybody had any feedback thoughts about their proposal, please drop it inside the discussion that will spur more conversation. + +00:50:00 + +Tyler Wright: The community and that's what we're looking to do. That'll help drive next steps there. So please ask any questions even if something seems like it may be a dumb question. Feel free to ask and you'll get clarity from the netted out team in peace and we can continue to push that process forward and as always thanks for the updates Max and all the things that you and the cloud most slash console team are working on there's a couple other teams that again we're going to continue to push to join these meetings. So for those that are new Feel free to reach out to me. If you just want to be added to the meeting every month. Again. There's a number of teams based all over the world that are solving a number of problems. So this is usually a very Lively special interest group on a monthly basis. + +Tyler Wright: All right, I hope everyone has a great rest of their day. Unless anybody has anything else. I'm gonna let you go and we'll talk more online. Cool. Thank you everyone. + +lukas meilhammer: Thank you everyone the pleasure. + +Tyler Wright: Bye-bye. + +Anil Murty: it's kind of + +Piyush Choudhary: more than about + +Meeting ended after 01:04:16 👋 + diff --git a/wg-provider-audit/README.md b/wg-provider-audit/README.md index 74f68390..a2389f54 100644 --- a/wg-provider-audit/README.md +++ b/wg-provider-audit/README.md @@ -15,8 +15,8 @@ This working group is for developing an open source auditor tool for providers. | Meeting | Time | Notes | Transcript | Recording | --- | --- | --- | --- | --- | | #1 | Wed, Feb 21, 2024 10:00 AM PT (Pacific Time) | [Link](https://github.com/akash-network/community/blob/main/wg-provider-audit/001-2024-02-21.md) | [Link](https://github.com/akash-network/community/blob/main/wg-provider-audit/001-2024-02-21.md#transcript) | Coming Soon -| #2 | Wed, May 1st, 2024 09:00 AM PT (Pacific Time) | | | -| #3 | Wed, June 19, 2024 10:00 AM PT (Pacific Time) | | | +| #2 | Wed, May 1st, 2024 09:00 AM PT (Pacific Time) | No meeting | No meeting | No Meeting +| #3 | Thursday, June 13, 2024 10:00 AM PT (Pacific Time) | | | | #4 | Wed, July 17, 2024 10:00 AM PT (Pacific Time) | | |