diff --git a/sig-clients/meetings/008-2023-08-16.md b/sig-clients/meetings/008-2023-08-16.md
new file mode 100644
index 00000000..af210067
--- /dev/null
+++ b/sig-clients/meetings/008-2023-08-16.md
@@ -0,0 +1,560 @@
+# Akash Network - Clients Special Interest Group (SIG) - Meeting #8
+
+## Agenda
+- Updates on Cloudmos and Console
+- Demo Presentation from Source Network
+- Discussions on Clients Libraries
+
+## Meeting Details
+- Date: Thursday, August 16, 2023
+- Time: 9:00 AM PT (Pacific Time)
+- [Recording](https://56oikxuxws6ewboifrwsgwvq2da4gzrg23sx7u5bldufgusgpc5q.arweave.net/75yFXpe0vEsFyCxtI1qw0MHDZibW5X_ToVjoU1JGeLs)
+- [Transcript](#Transcript)
+
+## Participants
+- Addo Smajic
+- Andrew Gnatyuk
+- Anil Murty
+- Anish Maddipoti
+- Boz Menzalji
+- Calvin Harris
+- Carlos A
+- Christopher Berry
+- Damir Simpovic
+- Denis Lelic
+- Emmanuel Arthur
+- envy | Neta DAO
+- Fred Carle
+- Greg Osuri
+- Ishan Dhanani
+- Joao Luna
+- John-Alan Simmons
+- Joseph Tary
+- Julius Haus
+- Keenan Nemetz
+- Mark Ntim
+- Max
+- Maxime Cyr
+- Rodrigo Rochin
+- Shahzad Lone
+- Thomas Montfort
+- Tyler Wright
+- Tzu-Jung Lee
+
+
+# Meeting Notes
+
+
+### Updates on Cloudmos
+
+#### Max's Update:
+- Max expresses gratitude for support on [ the proposal](https://www.mintscan.io/akash/proposals/216), noting high participation.
+- Ongoing work on the CloudMos Interface and Tools, now open source.
+- Integration of AkashJS library for SDL parsing and manifest creation.
+- Anyone can use the indexer acquired by the community with documentation in the repository.
+- Announcement of an upcoming release: the AuthZ manager for managing authorizations.
+- Mention of plans to implement authorizations that can pay for deployments fees, inspired by Omniflix platform.
+- The plan is to provide free trials for users without tokens.
+- Mentioned that they have "no public roadmap," but they work closely with Anil on things to improve and build. 
+
+#### Tyler Wright
+- CloudMos and Console teams supporting a sandbox environment for testing network changes.
+- Two bugs identified and resolved through testing.
+- Community involvement in testing to ensure client functionality on mainnet.
+
+#### Anil Murty
+- Anil addresses a question from Julius and discusses synergies between the two teams(Cloudmos team and Console team).
+- Emphasizes that Cloudmos is now open source on Akash Network.
+- Highlights the product strategy aimed at leveraging synergies between the two teams and finding common features.
+- Discusses the importance of finding common features between both platforms and reusing code where possible.
+- Details the plan to identify common features and integrate them into AkashJS.
+- Mentions that all the document he is presenting will be available in the Sig-Clients within the Akash Network's community repository.
+- Discusses plans to clean up the roadmap and make it accessible, mentioning the Cloudmos deploy folder as a potential location for it.
+- Highlights recent developments by the Cloudmos team, specifically UI for viewing and setting attributes for providers.
+- Explains that providers connected to the wallet can now see and modify attributes they've set.
+- Highlights that clients can view the attributes they're assigned.
+- Clarifies that the intention is not to become a provider management platform but to test the [proposed schema](https://github.com/akash-network/cloudmos/blob/main/config/provider-attributes.md) for compatibility with providers before wider implementation.
+- Highlights that attributes implemented by the Cloudmos team to help clients query and filter providers based on specific criteria, improved the client experience.
+- Highlights that attributes implemented by the Cloudmos team addresses an issue with the SSH shell interface in Cloudmos, which previously experienced frequent timeouts. 
+
+### Updates on Console
+
+#### Anil Murty
+- Anil Murty mentions that some bugs were reported in the Cloudmos channel on Discord during the Testnet phase. Fixes were made, including addressing an issue related to persistent storage.
+- Anil Murty introduces a new landing page for Console 2.0 but clarifies that it's an interim solution until the new experience is determined.
+- Anil Murty provides information about the Agora Labs team who are starting a project to build upon Console.
+- Discusses their aim to create a user-friendly experience for AI developers.
+- Mentions the team's plan to work on a prototype for an upcoming hackathon.
+### Discussions on Client Libraries
+- Joao Luna: Provides an update on the client libraries. The Java library is open source, while the GO library is still closed source and under discussion. 
+- Greg Osuri: Expresses the need for a GO library for his client project and is open to either building it himself or receiving access to one.
+- Joao Luna: Agrees to discuss the GO library further within the working group for clients.
+- Greg Osuri: Discusses the funding aspect and suggests putting out a usable version of the library first. He mentions that he will work on it regardless of funding but would need funding for further development.
+- Joao Luna: Shares insights into their approach in building the library and their need for a clients library to interact with the Akash network.
+### Demo Presentation from Source Network
+- Tyler Wright: Introduces the [Source.Network](https://source.network/) team and gives them the opportunity to demo their decentralized solution.
+- Addo Smajic briefly explains the nature of their decentralized database solution.
+- John-Alan begins the demo of their database solution, [DeferDB](https://github.com/sourcenetwork/defradb). Launches an instance with an in-memory store, explaining the use of schemas and Graphql for interacting with the database.
+- Greg asked for a reminder about what a CID is.
+- John-Alan explained that CID comes from the IPLD data model, which powers IPFS and Filecoin, and it stands for "content identifier."
+- John-Alan mentioned the importance of the Merkel CRDT system for building decentralized databases.
+- John-Alan mentioned the need for developers to spin up nodes on Akash to act as replication endpoints for users going through applications.
+- Greg expressed enthusiasm for the Source project and its alignment with Akash's vision.
+- Discussed the idea of having managed services, like databases, that can be scheduled with containers.
+- Greg inquired about the consistency model supported.
+- John-Alan explained that it falls under the "strong eventually consistent model," where all clients see the same state as long as they are on the same version.
+- Greg Osuri inquired about the number of verifications required for eventual consistency, but the exact number wasn't provided.
+- Greg asked about ORM (Object-Relational Mapping) support.
+- John-Alan mentioned that GraphQL provides a native ORM-like system.
+- John-Alan highlighted Akash's ability to allow users to spin up VMs easily with a private key, reducing the complexity of server management.
+- A question was raised about how authentication to the database works.
+- John-Alan explained that they have a Global Authorization service within their Cosmos APP Chain.
+
+### Other matters Discussed
+- Thomas Montfort provides a high-level overview of their projec and background as AI/ML engineers.
+- Highlights their interest in Akash due to the GPU testnet and the support for GPU resources.
+- Expresses the challenges in deploying and fine-tuning models due to resource availability.
+- Outlines their immediate plan for the HACKATHON: creating an interface for fine-tuning models and automating deployment.
+- Mention of handling model checkpointing and dealing with providers.
+
+#### Model checkpointing
+- Anil Murty: Discusses the topic of checkpoints and shares that they are in talks with another team working on Lower Training, a technique that involves training large AI models on a smaller subset of parameters. They are exploring the possibility of enabling this training on Akash once the mainnet is launched.
+- Anil Murty: Highlights that the training aspect of AI models can be exorbitantly expensive, and the Lower Training solution could be a cost-effective alternative.
+- Anil mentions that he is working with the team on a proposal that will be published on GitHub discussions in the next few days.
+
+## Action Items
+- Cloudmos team to release new features and additional functionalities.
+- Anil Murty to publish a high-level roadmap for Cloudmos and Console on GitHub.
+- Anil to share presentation documnent in sig-client repository
+- Luna to continue discusions with Greg on Go library development in sig-client discord channel
+- Tyler Wright to check if the Source Network team has plans for future sessions.
+- John-Alan Simmons to prepare a demo for a larger audience.
+
+# **Transcript**
+
+_This editable transcript was computer generated and might contain errors. People can also change the text after it was created._
+
+Tyler Wright: All Welcome everybody. To sync clients monthly meeting number eight.
+
+Tyler Wright: At these special interest groups for clients, it's an open discussion with folks that are running clients on the Akash network, you can look at the Readme where you can see a number of those clients in there that again, usually end. Some of these meetings, we have a fairly full agenda as mentioned today. I know there's been a lot of added the last dig clients meeting there was some talk about some collaboration going on between cloud most and console. Obviously, over the last month, there's been a lot of progress made there including again proposal that just passed yesterday for community funding for the cloud most acquisition. So again, shout out to all that helped out there. We want to get some updates from a number of folks that are running clients and then following those updates from Team Tech Cloud, Most Console, and others, that might be present. I see Luna's present that does, again clients
+
+Tyler Wright: there will be a presentation from Source Network. We'll be giving a demo and some of the work that they've been doing and how they'll be leveraging a class network. So, without further ado, I'll hand it over to. We'll start off with, I guess Max and Cloud most, because I know there's been a lot going on, and then from there, we'll go to Anil and we can continue on down the list of updates from clients. So Maxi want to start us off.
+
+Max: Sure, everyone. First of all, yeah, thanks for the support on the proposal. I think we broke your record. In terms of the participation from the proposal. So yeah, that's awesome. So yeah, we just created the repo for the open source people.
+
+Max: share it in the chat. yeah. Right now I also created the first pull request With the changes.
+
+Max: Give me a second. Yes. Either changes of icast. integrated In the Deploy Web previously, we had the custom code, To handle all the SDL parsing, and Creating the manifest and the version. But now it uses the ICASTGS library.
+
+Max: We have a lot of stuff. Going on to keep working on the CLOUDNOS. Interface and Tools. Right. Now since it's open source, anyone can use the indexer that we built and I'll acquired by the community. There's a bunch of documentation in the repo. That you can follow to run your own indexer. And you can play with it. There's a ton of stuff possible to do. With the indexer in the database. So, if you have any questions on how to run it, And what to do with it. Don't hesitate, I'm always the available on the discord. So yeah, in terms of the changes and stuff, we have a
+
+Max: not public roadmap but we work closely with Anil on things to improve and Things to build will be releasing. This week, the Adzi manager. That will Managing Your Authorizations in. And a small dashboard where you see all the list authorized addresses that you have. Authorized to spend on your behalf. Which will make it a lot easier for. Enabling free trials or just making someone or pay for someone else's deployments. And we also plan on A very shortly inspired by the Netflix platform. They did this.
+
+Max: When you meant an NFD and OMNIFEX they also pay for the fees with the authorization. So We'll add this as well, so you can just authorize an address. To pay for the deployments and also for your fee. So, It's this great for giving free trials for new users, they don't have any tokens. They can just get the authorization and try to deploy something for free, basically. Well paid by the user, that will grant the fees and the deployment. And a lot more stuff to come. But yeah, that's pretty much it for now can follow along and start a repo and Watch for pull requests and stuff, we're going to be pretty active on it.
+
+00:05:00
+
+Tyler Wright: Thank you, Max. I see Julius had a question about. Are you all going to be switching to Akash JS in the future and just some plans there and does anyone have any other questions that they want to either throw in the chat, or others? Go ahead and you
+
+Anil Murty: I can actually answer that when I speak. And I can share a little bit of information on that as well.
+
+Tyler Wright: All right, so we'll come back to that question again. One of the things that the cloud most team is doing as well. The console team they are supporting sandbox environment for a number of folks on the insiders vanguards community. Are testing the network changes through the cost, GPU AI Incentivized Test Center. There was two bugs that came out of that test net, those bugs again had since been resolved have been tested by the Overclock Labs team and I've been tested by the community using the CLI, We want to make sure that again those bugs are not still having any issues or excuse me. The clients rather are able to do what they need to do when we move on to mainnet. So, again, that's another thing that the cloud most team is working on. Does anyone have any questions for the cloud most team before we move on to an update from Anil on all the efforts from the consulting?
+
+Tyler Wright: I'll hand over to Anil to give us update on console.
+
+Anil Murty: Thanks and I'll talk a little bit about cloud mostly console both because, The clubhouse team in me, and Joe have been working pretty closely. Leading up to submitting the proposal and stuff. So what I'll do is I'll just share my screen.
+
+Anil Murty: so yeah, as Max said, Promptly. As soon as the proposal passed that most people is open source and kash network now. So check it out if you haven't already.
+
+Anil Murty: The first answer the question that Julius asked, I think so prior to us, sort of writing the proposal and stuff and going through the voting on this and putting it out there, We actually talked through how we could have synergies between the two teams that was as well as console. So essentially So, a little about this as part of the Proposal that we put on discussion as well. And there's a product strategy, dog, regiment read, but this is fairly high level. And I understand there is really high level but we kind of want to do not go super detailed on this one. But we have been working together to figure out what areas where
+
+Anil Murty: There are things that are common feature between the two clients. So, I called out in the strategy. Doc essentially cloudmose is going to continue focusing on battery, whether they've done really well and then console is going to pivot and focus on trying to go after web to users and make a product. That's really attractive to them specifically initially, at least AI developers who have to. But all that said, there's a lot of features for a deployment clients that apply to both types of users. And it's in our best interest to make sure that the code that is used in both of these products is reusable so that we're not repeating and redoing things in both the platforms going forward.
+
+Anil Murty: luckily we have a library called Akashias and this is also something that is publicly available and is open source. So if you haven't looked at it akashiases right there, And so, the idea is that we would try to find common features that exist between the two platforms. And if they don't already exist within akashias, we would pour that code into akashiya. So if it's a feature that's implemented in Cloud, most that doesn't exist in console today and Cloudmoms has the better cloud noses implementation. We get into akashias so it's available to both the products as well as any other clients that print that to choose to Similarly, on the console side as well. So just to answer the other questions that came in I think, from Anish or Ishan.
+
+Anil Murty: this document is right now, just a Google docs that I shared between us in the clubhouse team. I am good to work on getting this into Github in the next couple of days. So it'll be available to all of you. And then along with this, we're also going to publish work on publishing a high level road map of what That was going to be working on and all of this is going to go into the C clients. So again for people, the people that are not familiar with this or joining us for the first time, if you go to Akash network here, one of the repos here is the community people. And then under community, we have various special interest groups and one groups. So one of them is sick clients were group, which is what this call is related to. And inside this, there are the various clients projects that are being worked on that. We typically talk about on this call. So what we're going to do is once
+
+00:10:00
+
+Anil Murty: Clean up the roadmap, we'll put it inside. The cloud, most deploy folder here for anybody to look at and then we'll figure out, somewhere maybe we can put the table that I showed you in the word doc as well. I'll take a look at Cool. That's the general updates. I want to talk about in terms of answering that question. As we're talking about Cloud, most places console, a couple of things that I think Max probably didn't mention it simple, but I would love to talk about in terms of things that cloud, most team ship last week is number one, the shift UI for being able to view and set attributes for providers. So essentially if you're a provider that's connected to the wallet and you go to cloud, most now you can actually see Attributes that you've set for your provider and you can change those values for those keys as well.
+
+Anil Murty: And then if you're a client, you can actually view all the attributes that you're seeing. I posted a note about this after Max posted on the WG Working group for provider attributes and discord and basically, I just want to clarify that. There's no intention here to become a provider management platform. That is, creator guys are doing a really good job of managing that from a UI perspective. The intention of doing this was few Lee to do two things. Number one, be able to test if the schema that is proposed, earlier in a word document and now in this link that Max is just shared testing, if that schema is going to work with providers. And if it's not gonna have testing if it's gonna have some issues with certain the way things and named or, the way the values are created. Before we roll this out and mass or scale it out to all the providers.
+
+Anil Murty: That's kind of number one and then the second is This relates very much to the client experience because today, when you're deploying on Akash, regardless of which clients you're using one of the challenges you have, is that there is no way for you to query all the list of providers and say, I only want to see providers being able to say things like only show me providers that are located in a certain region of North America or, providers that offer a certain type of capability. Whether it be a support or something else. So the idea is to have attributes that can help us query as well as filter on providers based on those requests that the clients might have. And so as we scale the number of providers and Akash, let's say we get into the hundreds of providers, somebody should be able to apply these filters and just look at a list of 10 or 20 providers and decide. What's working for them. That's kind of the intention So provide attributes implemented by the cloud. Most team amazing job. Check it out on
+
+Anil Murty: Our most deploy, if you haven't. And then the second one, which is sort of smaller, but it's something that is bothered me as well as hopefully many others reviews cloud. Most in the past, is If you've used SSH the shell interface on cloud. Most
+
+Anil Murty: You may have noticed that it often feels like it's timing out so you enter a couple of commands in the container and then come back to it a few minutes later and no longer have access to the shell so it almost Emmanuel fix to that as well. So now it should be such that, you can navigate away from the tab that has a shall Come back to it, initials still should have access to the container and this is really important particularly for people that are doing, AI Fine tuning or inference, where you want to have access to the shell. Copy a bunch of parameters, or download a data set into your container, that's running the model and losing access to it and constantly having to open the share this kind of pain point. So, great job on that as well. Cloud Nose team,
+
+Anil Murty: On the console side, there was a couple of bugs that were mentioned in the Cloudmos, channel on discord as people went through the Testnet. So I believe you has made the fixes for those. One of the big ones was an issue around persistent storage. One of the things that I do with console is we're trying to not worry about bugs that, are going to be purely on the Vetri side. Because going forward, Moses going to be the tool that will be used for that. But if it's seeing that they're going to impact, even the Web 2 version of this console, we're going to go fix them. So that's why you might see. and things getting prioritized a little bit lower, the console side but it's in a staging environment that we host internally
+
+00:15:00
+
+Anil Murty: and then we also have a new landing page that we were working on prior to this acquisition and, changing the direction. So it looks something like this. So it push this out the production and, this is not like something that is going to be the new landing page for Console 2.0 or something. It's just being interim until we figure out the new experience. So that's kind of the main updates on the cloud nose and console side. I do have another thing I want to talk about in terms of a conversation that we have with another team, but I want to pause here. See if there's any questions on Cloud Mods or console. To check the chat really quickly.
+
+Anil Murty: So I'll take a couple of minutes Tyler just to talk about it. One more thing since I have the mic right now. So I think they should be folks from Agora Labs on this call as well. ishan's on. I think maybe a nation thomasville I see a Thomas from Montfort's Maybe that's down. Hey, don't so good.
+
+Thomas Montfort: How's it going?
+
+Anil Murty: Thanks for helping on short notice. So I just think these guys five minutes before the call and asked if they can join. So they thanks for joining So Tom Ishan in Anish, they're kind of starting to work on a project of their own and hoping to build upon console going forward. we talked a little bit last week, and, they have some ideas around building, something that is very similar to the strategy that was outlined as part of the Cloud Nose acquisition For building something that is a good user experience for AI developers and doesn't have a very very native interface to it. So I just kind of wanted to put that out here and I was telling the Agora Lab team that
+
+Anil Murty: Earlier this year, we made the decision to go 100% open source and all the things that we do are done in the open as part of these things. So, once you guys figure out, what the user experience is going to be and will work together on that. We should publish respect for it here. and installing that, but in the most immediate term, I think the plan for the Google team is to work on a prototype for the upcoming HACKATHON. So that's kind of the initial milestone. So with that, I'll kind of see if you guys want to share anything more than that ishan or Thomas or anybody.
+
+Thomas Montfort: Yeah, I can just give a brief high level overview but yeah, we really were AI ML, engineers by trade. And what really struck us with Akash was the GPU testnet and this support toward GPU Forward is insanely huge if you're trying to deploy or fine-tune a model right now it's just insanely Very hard to get to resources so that was really what caught us And so we've just been diving into the community just kind of seeing what is being built. And we notice kind of, with the very web to focused, Emma engineers. They're not going to want to have to get tokens deal with wallets all of that. But are more immediate plan for this. HACKATHON is really building out an interface for fine-tuning models. So you would be able to say
+
+Thomas Montfort: Here's my data set stored in S3 bucket or somewhere else. Here is a preset model that we have, maybe it's Lama 2 and then we're going to automate the deployment to the network itself. So that's kind of high level our plan and obviously that brings into question or thought like checkpointing the model How do you deal with providers shopping off? So yeah, that's just high level. What we're planning on building out.
+
+Anil Murty: Thanks for that quick introduction and Appreciate you guys jumping on in short notice again. Yeah, if there's any questions on that. I mean, it's super early. They haven't really figured out what they're going to build, but the good thing is, There's a hackathon coming up. So hopefully we'll see something soon.
+
+Anil Murty: on the topic of checkpoints. The other thing that might be worth sharing is more of a heads up. Is We've also been talking with another team that is building kind of for a solution for being able to
+
+Anil Murty: Do something called Lower Training which is low rank adaptation of large language, models or large models In general. The idea with that is you take an existing, really large AI model and rather than retaining it on the full set of parameters that it was very train it on a subset of those parameters with a smaller data set. And the idea is that if you are looking to improve the model by tuning it on a smaller data set, it lets you achieve that without having to use the massive amount of resources that was used previously playing the model to train the base foundation model. So we're talking to this team and we're trying to figure out if there's a way for us to enable them.
+
+00:20:00
+
+Anil Murty: To be able to do this training on Akash when we get to the main So once we launched me in it, we'll look at seeing if we can enable them to train on our caution. It'll be a good example of being able to achieve that use case on Akash as well.
+
+Anil Murty: Because if you think about it, as part of the recent test net, we demonstrated a couple of things. We demonstrated being able to do inference. There's a bunch of people that build applications for infants and random And then we have had examples of people taking an existing model and, fine, tuning it with a smaller data set, or using something like envy chain to embed together, two different Llms or two different models together and utilize them. So that's kind of in the second camp of not just, running inference. But doing a little bit with the model, the training aspect of it is something that is exorbitantly expensive to do in general, if you wanted to train something and you need course you several millions, if not seven hundred millions of dollars to do. But this idea of using Luna is something that we think could get is
+
+Anil Murty: A good demonstrable example of training a model without actually spending that much of money. So the plan is that I'm working with that team to write up a proposal and I'll be publishing this on github discussions in the next few days. so that's just another update. I want to share as part of this. I think that's pretty much all the updates I had on my side for now.
+
+Tyler Wright: Does anyone have any questions about anything il Just talked about
+
+Tyler Wright: Is again pleasure to meet the new team. If you need anything, please feel free to reach out and SID clients and we can make sure you have everything you need. Cool. I do want to make sure there's enough time for the source network team, I know that we'll still be working on clients libraries. I know some of those efforts are still in discussion. Do you have any updates that you want to share in the work that you've been doing? I want to make sure that I give you some time if you need it.
+
+Joao Luna: Yeah, regarding progress on the libraries. As far as I know, I think the Java one is open source right now. The goal one is the one that's still close source. We are just using it internally to develop our own product and it's still up for discussion. So, I think that's the effort. That's coming in and it's really just waiting for the mainnet to have a stable release. We are currently just using the release candidates on test and sandbox.
+
+Greg Osuri: Hey, I'm starting to work on a clients and I need a library. Either I build it or you give me access to it. So it's gonna happen. Either way. so I would prefer you give me access good because I would rather build on your work versus, I'm going to do it, but I do need to go library. I think it's important that we have a go library before a Java library, considering not to maybe we'll use Java or, other stuff we're on Akash. But look go, I wish I could.
+
+Joao Luna: Yeah. that's something we can definitely discuss further on the working group for the clients,
+
+Greg Osuri: Yeah, and the funding I think it's a clients, It's a library. I mean, Give me two days. I'll give out a version, I think we should put something out there first And grow right a roadmap. As I want to develop this library and get funding for that, right? Because I think it's too small of an initiative. Not to discount your work, but I'm talking to me.
+
+Joao Luna: So yeah.
+
+Greg Osuri: Standpoint of I built the library So I would do it myself if otherwise, I think it's too small often effort to warrant an entire funding proposal on shame but if you put something out there that is usable to certain point and of course you're going to need funding to continue working on it and then we can write a better diggers land, or roadmap and a proposal for that and public, So I'm going to do it regardless of funding, right? You need to be a library to the product then. That sufficient tomorrow 880 toward my involvement without having to actually get paid. so that's just my take
+
+00:25:00
+
+Joao Luna: Of course. Yeah, that's something we can discuss. It was exactly our approach. so you can't,…
+
+Greg Osuri: Okay.
+
+Joao Luna: you kind of like a real library but already existing clients we've just found out. We're just two. The troublesome to use. So we kind of designed something that we as developers would like to use as a library and that's kind of what we built and where it was born from the only caveat it was that it was
+
+Joao Luna: Implemented according to the specification that was built at the start of the discussions on clients libraries. So that's kind of like that the twist that we had with apart from that this is something that we already were building because as you mentioned that you are requiring as well. We needed a clients library to start working on backend services and go there, interact with the cash network. So to make our product possible.
+
+Greg Osuri: Yep.
+
+Joao Luna: Yeah, it's fine. It's something that we can definitely further discuss on the working group.
+
+Greg Osuri: Sounds good. Thank you.
+
+Tyler Wright: Thank you, Greg's. Anyone have any other comments for Luna before I handed over to Source Network to give a demo.
+
+Tyler Wright: Before I do that, is there anybody from the Fleet team or sphere on team that wants to give a quick update?
+
+Tyler Wright: All right, if we'll get updates from those teams between meetings, I'll make sure they put something in the sick clients discord channel. Without further ado, thank you for all those updates from the teams and all the work that's going on. I do want to hand it over to the Source network team. I don't know if Ado Addo, I know that we've been talking in clients, as well as just generally in discord. They are a team that has been working on a decentralized solution that's going to be leveraging. Their cash network and I want to give them the opportunity to demo their solution in front of the group. So I'll hand over to Edo to get us going.
+
+Addo Smajic: Yeah, it's Addo. And …
+
+Tyler Wright: Of apologies.
+
+Addo Smajic: yeah, we have ours, I mean, I'll be quick. It's a decentralized peer-to-peer user-centric native, privacy preserving. No SQL database that we build from scratch to essentially a kind of plug fix that, issue with centralized databases that's plaguing. What three and kind of allow developers to bring true decentralization to their Applic. And I'll hand it off to John or CTO to kick off the demo.
+
+John-Alan Simmons: Adam, and ler. Thank Appreciate being here. You give us a chance to, show our stuff and hopefully gets really good feedback. So, let me go ahead and get my screen share going.
+
+John-Alan Simmons: And I think this is the one.
+
+John-Alan Simmons: Hey my various terminals and windows at the moment. Okay.
+
+Tyler Wright: Yes.
+
+John-Alan Simmons: Perfect enough the sides, Okay I can be a little bit open on some smaller screens. But I think we should be good to go. Alright so yeah, welcome we're building this presentation's database that I described. we're gonna go through today is really just kind of like what the mechanics and features and goals of this database is. And you know how it kind of affects the developers. What they really unlocks from an application perspective and kind of walked through some of the stuff that we're working on, obviously in relation to a cash and this ecosystem community and so, with that, we're gonna get started, we're just gonna get, a quick little instance, go into this database running on my local machine. I'll kind of go through a relatively soft and slow progression. And depending on what we get through, we get through real quick. Tyler, approximately. How long would we have for the total demo?
+
+John-Alan Simmons: Just want to make sure.
+
+Tyler Wright: We've gotten to the rest of the agenda. So we I think some people might have a hard stop in 29 minutes so we have another 29 minutes,…
+
+John-Alan Simmons: Perfect.
+
+Tyler Wright: I want to lay some clients at room for questions as well.
+
+John-Alan Simmons: That's most certainly. All And feel free to anyone, if there's anything that pops up there at the demo, feel free to pause and ask a question there, to wait for everything till the end. All right, so I'm gonna go ahead and run a instance of defer DB on my local machine. As I said, We're gonna start it up and I'm just gonna tell it to use the in memory store, I usually do this for demo purposes, just in case. I'm making a mistake. we can cook the delete the process and all the data will go away. But we do support our in memory store, Obviously, it's more performant, but, does not persistence. And then of course, we do support file back persistence and we're working towards, connecting that to long term storage providers. Are we see a storage day? Kind of, that whole, because of some things about going. But at the moment, we're just gonna start this in memory store.
+
+00:30:00
+
+John-Alan Simmons: So it's gonna go ahead and just obviously put out some debug information at the moment. We don't care about anything but there's some interesting bits and pieces that will come back to you. In this notably some of the lid, pdp stuff that we have going on.
+
+John-Alan Simmons: And first and foremost, this is a No sequel, kind of document store database, but it is fully schemaed. So we're gonna go ahead and add a schema, from a schema definition perspective and really whole query language perspective, our entire database is based and built on on and around Graphql. we were deciding pretty early on how we wanted to have developers interact with and build in the system and what kind of query mechanics we're using. And, because of the way we're doing kind of this, no sequel document store process. And, we can really go down the SQL route. We didn't want to go down a route of creating our own custom query language because, that felt like that just was a huge burden on developers to learn. And so, Graphql kind of at the time was a very early thing as far as being integrated natively into databases, obviously they've been around just doing API things for quite some time. But, we were think the third database to actually start kind of like supporting Graphql as a first.
+
+John-Alan Simmons: This isn't on a database level, which was actually unlock, a really interesting, set of things that you can do with graphql, It solves a bunch of kind of traditional problems with the graphql but we'll get into all that. And so, we do support a bunch of kind of pretty interesting and rich models and relational structures and everything. You can see here this example kind of is a combination of one to ones. And one to many's the database managed a relationships for you. So you don't have to do a lot of the foreign key and primary key, definitions inscriptions to figure out all the making. Sure those lineup, if you just specify the type and the fields, then the database, if you so choose we'll basically manage that whole relationship for you.
+
+John-Alan Simmons: But in our kind of quick little demo here for the next 30 minutes, we'll use the straightforward one. If there's time we'll kind of transition to a little more complicated one and see some of the differences there. so straightforward. this particular user struct is gonna have four fields name age, verified and points and it's gonna have the respective types of string and Boolean and float. And again, this is just the graphql's way, defining types. So we'll go ahead and get this executed on the database. We'll do definitely be Clients schema ad and then we'll specify Pack F, which is going to the house to read from a file. And we're gonna read from that exact. So unsurprisingly that did exactly what we kind of expected it to do, and
+
+John-Alan Simmons: This is our playground. I actually comes bundled with our instance of the database and it's really just an easy way to kind of get going, sometimes you might need your own, native Graphql, clients or whoever it is. So if you want to just kind of get going and run it up, and testing and playing around with the database. This basically ships with the database, that's the bet. So you don't have to go get your own l clients. the really beautiful thing about Graphql and I know some of you may not but it is as you just saw a fully strongly typed and fully schemid language and so because of that you actually get a really kind of rich collection of
+
+John-Alan Simmons: Pipes that are actually generated behind the scenes and that's part of our database. Some of us just kind of native to graphql. which means that we get this kind of really great automatically, generated introspection APIs that go through all the types and all the systems that are defined. You can go through all the kinds of queries and because of that we actually get really fun introspection. I fired quickly just kind of the first second say we were going to create that query from Scratch mutation. Create user. we get the automatically done there. This we have to kind of do as a JSON blob. and stopping to kind of see in real time that
+
+John-Alan Simmons: action happening. And this is just a benefit of kind of graphql again to use battle ecosystem. Here, we get to introspect on the various fields key, which is the the Doc ID system, and we'll get into what that means in a moments name age. Verified and points.
+
+00:35:00
+
+John-Alan Simmons: So same for as one here. I just want to kind of get you guys a chance to see that. you get that kind of like automatic IntelliSense right off the bat. We'd have to do anything to support that. All right, perfect. So we'll go back and run our original query. Because I know what to expect here. And we get back. Just this data payload, which is the response back from the database And the only thing we're actually asking for at the moment, is this ducky but we'll get into kind of what that means and how that really starts leveraging the powers of both mutable and immutable data and how that can be really really great for crustless verification of data state.
+
+John-Alan Simmons: So now we'll just go and run a general query. The name here doesn't actually matter. This is just a way to name a query and I like to do it because it means I get my little names on the top pads here, but that's just Handle thing to do. And so we're gonna go ahead and run again just like the general user query which is just the way to interact with the user collection as a whole. And so this user collection is what we define is our type and has our fields and it's the thing. We watched it, too. I'll add all the other bits and pieces here. and you can specify these in any order. Graphql is going to give you only exactly what you asked for. So if I don't ask for a certain fields, it's just not going to give it.
+
+John-Alan Simmons: And there we go unsurprisingly. Yet again, we're getting back to the exact kind of state that we ended up that we put into the database. You'll see though. One interesting thing that is that we didn't actually specify key parameter this Doc key. And so, again, like I said, behind the scenes, when you insert this type into the database, there's actually a whole series of codegen and scaffolding that's generated and all these, associated types so that we can do filtering type, safe, query and type safe, selection sets, you're able to do a lot of things and the database kind of leverages a lot of that for you. And so going through this, we'd have to go and find the key system, it's just there.
+
+John-Alan Simmons: Speaking of the ducky this is where the first kind of interesting mechanics and notions of how this database and what the underlying mechanics of this whole system is. And so I mentioned kind of we wanted to marry the concepts of both mutable and immutable data. And if any of you have worked with just the Ipfs API, as a whole and trying to connect build out a system using nothing but that Ipfs API and the Ipld data model, you'll know that it quickly becomes kind of
+
+John-Alan Simmons: Really exhausting and difficult task because you're always trying to stay up to date with whatever the current CID is of your state. And you have to do kind of a lot of effort to make sure that that stain and sync and potentially maybe you use an Ipns to do that. And you kind of have this static identifier that links to a mutable state Ipns has its own issues of slow query resolution and the records only stay around for 24 hours and it's global Dhd. And so for us the Doc key here it's kind of like our approach to having the familiar developer experience when it comes the databases of a reliable primary key that you can interact with but you still get all the benefits of that rich immutable, I appeal dhana is really just a CID
+
+John-Alan Simmons: In a UUID and is consistent and static for the entire lifespan of this document. If it gets updated and synchronized over the peer network and forked and branched and we'll get into what all that means. This Doc key will remain consistent and so if you take that key to any peer, you'll be able to resolve the same document and then verify that state and we'll see what that actually looks like. So again unsurizing we're getting back exactly what we're kind of expecting to get back. But I mentioned that we have this branch in your forking model and so if we were to go ahead and send an update to that particular document, using the update user mutation, we're going to specify just make sure that's the exact Keep you want? Yes, E5.
+
+John-Alan Simmons: So we're gonna send an update to just update the age of this particular user to 123 because apparently the most person in the world, we're gonna send the command again, unsurprisingly, we're getting back. Just the thing that we care about and if we go back to get, it's going to unsparsely, give us back these things. And I'm gonna say unsurprisingly enough because there's something that there's aspects of database development and as developers and consumers of the database. there are things that you just don't want to have to worry about bleeding edge technology, especially when it comes to databases because you want that ance. You want that, guarantee that some things are gonna be just there and working and the way you expect it. and so this database as a whole has kind of two.
+
+00:40:00
+
+John-Alan Simmons: Allergies are too personas and we like to say, looking from the outside in, it should just feel like a database because foundationally is we've built this from Using the ipld data model, using lip P2P using Merkel, crdts using all these verifiable and structures, but at the end of the day, You should just care about that. It's a database and it's gonna fulfill your nudes as a database. if you want to start peeling back the curtains and actually understanding, how it's exposed this information and what that means for developers their applications and their end users. If we go to this commit step, and we'll go and get rid of this for a second. And we run this word. Yep.
+
+John-Alan Simmons: Yep. Absolutely. Yeah, It's a somewhat long-winded way to kind of work up to that because, all these factors go into what it means to be kind of a decentralized peer database and then how that mechanics and then why that decentralization in that peer become so important. When deploying it into, compute domains like a cash, so, kind of a somewhat long-winded explanation, but I will absolutely make sure that gets cemented. So, we're gonna go ahead and to run this commits API.
+
+Greg Osuri: So what's a CID remind me again?
+
+John-Alan Simmons: As so apologies CID comes from the Ipld data model which is the system that they're underlined data model, the powers IPFS and file coin and all that great stuff and it stands for a content identifier.
+
+Greg Osuri: Yeah, thanks.
+
+John-Alan Simmons: And so it's a way to, use IDs that represent the data rather than
+
+John-Alan Simmons: Everett And so using the database, this is how you feel like it should work but actually under the hood, this is actually how data is Data is interact with and data is built upon and this is where our Merkel CRDT system comes into play, which is a foundational requirement if you wanted to build out a, period of decentralized database because you want that ability to have that peer-to-peer because you want that ability, you have peers interact with synchronize in real time In first Merkel's crites, which is kind of this combination of using a Merkel clock from the IPLT data model embedded with a CRDT mechanic and a CRDT stands for free replicated data type. And with these conflicts, we replicated tips, you are able to have these really rich levels of collaboration without complicated or expensive global consensus.
+
+Greg Osuri: What kind of consistency does it support?
+
+John-Alan Simmons: So this falls under the strong eventually, consistent model, all clients will see the same state provided that they're on the same version. because each version is tagged and identify and it's structured as a graph or a DAG, then different clients that might see. Different states is only because they're just either further behind or further ahead in that day and…
+
+Greg Osuri: Are there any options for immediate consistency?
+
+John-Alan Simmons: And so, you're able to anchor these commits to our protocol and we have a Cosmo cell one. They're appchain that we're working on and so the first commit of any document gets anchored it automatically, and then from there on, you can kind of start. We'll be allowing to shoes. What is the frequency of the interval of those anchoring commits? If you don't need that,…
+
+Greg Osuri: That.
+
+John-Alan Simmons: strong global consistency, you can do it at whatever kind of interval that's batch together. If you want that guarantee you can anchor every commit and that's really just the CID that gets anchored on chain at that point.
+
+Greg Osuri: That how many verifications before eventual, consistency?
+
+John-Alan Simmons: So happy verifications for eventual consistency. I'm not sure…
+
+Greg Osuri: Yeah. How many interactions in terms of having …
+
+John-Alan Simmons: if I understand the question.
+
+Greg Osuri: Carlos, how to consensus work? How do you? Know.
+
+John-Alan Simmons: Yes, yeah. So the great thing about, the way Merkel series work, is that we kind of get this notion that we call consistency without consensus. And so, it doesn't require this global Really, any coordination amongst peers, it only requires just, verifying the DAG. When you guys appear receive it as a
+
+Greg Osuri: And set up a wall of type. because,
+
+John-Alan Simmons: Exactly. Yeah, baby very much. Kind of represents that wall mechanic. And again, from this being an ipld tag, it kind of follows those same semantics as well as, being able to pull down states, especially in different sets from different peers. And again, it doesn't require this global consistency, unless you want that strong mechanic, And so everything that was really good question. And so what this information actually, represents isn't all that important other than the fact that we actually get to structure and represent this data in a much more rich way and we can't kind of what we get out of that is really quite powerful, first and foremost, we have all these mechanics to query and filter and aggregate. I won't get into that for the sake of time and again, want to get to be cash stuff, we have the ability to run really, really powerful and rich migration.
+
+00:45:00
+
+John-Alan Simmons: Patch systems that are powered by this really great project that we also wrote in Haus called Lens VM, which is a webassembly based directional migration or transformation engine. It doesn't matter what that means of the fact that we get decentralized scheme management and decentralized migrations, which turns out to be a huge pain, database migrations are complicated enough adding in, stronger vegetable consistency, and dags and decentralization. And peer-to-peer turn is no easy feat. So, we're able to build out a whole dedicated, migration system to allow different peers on different versions, to still be able to interact with synchronize and work with one another even in real time. Perfect. Thank…
+
+Greg Osuri: That's really cool that you're making migrations the first class feature. that's always a secondary feature and…
+
+John-Alan Simmons: Thank you. Exactly. That's
+
+Greg Osuri: it sucks, but what kind of oram support you looking at Besidesm
+
+John-Alan Simmons: What kind of orem? So I mean the great thing about the way Graphql works kind of out of the box is that it almost has this native ORM system in just wrapped up.
+
+Greg Osuri: Okay.
+
+Greg Osuri: Got it. Yeah. I'm still in this old SQL world,…
+
+John-Alan Simmons: Exactly. Yeah.
+
+Greg Osuri: but let's go. Thank you. All right.
+
+John-Alan Simmons: And so because of that, we get to actually type, Jen a lot of our clients and so you get this, really rich …
+
+Greg Osuri: Very slow.
+
+John-Alan Simmons: here parody with your code and then because of the way, Graphql describes documents. It kind of like you're application state and your application domain is gonna more closely match, your data state and your data domain, or…
+
+Greg Osuri: Thanks.
+
+John-Alan Simmons: am kind of client code gen. Mechanic means that you have Really close parody between the two, which is really, really Perfect. In terms of a cache Keenan,…
+
+Greg Osuri: Cool.
+
+John-Alan Simmons: do you have the link for the running instance? Or is it not live at the second moment? I keenan's, one of our devs on the
+
+Keenan Nemetz: I can find a link. Yeah what's more interesting is probably the deployment.
+
+John-Alan Simmons: All right, perfect. I'll start walking through stuff, which wanna do it?
+
+Keenan Nemetz: What's that in the chat right there?
+
+John-Alan Simmons: And I have that right here.
+
+John-Alan Simmons: And so, the really important thing for this database among all the important things that I've gone through and hadn't have fans to go through, is the deployment models about how and where and what capacity and in what trust and authority those deployment models require or necessitate
+
+John-Alan Simmons: And so really, that's a long way to say that the goal, this database is to be extremely flexible to kind of suit a lot of different application needs a lot of different user needs, a lot of different developer needs and so that means it deploys in a lot of different environments from, decentralized compute networks. Like a cache, which has been the first decentralized computer network that we've targeted for myriad of reasons. Obviously now, which needs to be described this crowd but, the same time we want to be able to deploy to, traditional environments and AWS and GCP. And so we actually have partnerships with both AWS and GCP to make that happen. This is being worked on to be deployed.
+
+John-Alan Simmons: The context of a webassembly deployment so you can actually bundle it directly inside your browser and run. alongside your application code, which becomes really important for you for guarantees around User-centricity. But, because of the way that the peer system works, you want to have this mechanic such that you can have as many nodes in as many different domains as possible effectively. And so for developers and for applications, for users, it becomes a really great trust model Knowing that, users can basically freely verify that they're pushing their own, local database state that's being generated in browser via webassembly and that's being pushed to another peer and all peers are just, equal databases in the network that's being pushed to appear that's being run on a compute network where you can publicly verify that it's running the state and it's running the images and it's running, the things the way you expected to because obviously when it comes to again this relationship of trust
+
+John-Alan Simmons: Even in the decentralized world, there's always going to be that establishment of who are we trusting and what is the authority. And for us it was a simple no-brainer that we foundationally needed to support a platform like a cash to be able to have developers spin up those nodes so that they can act as replication endpoints for users as they go through applications and generate state and generate data. yeah. So this is the cash deployment. I wouldn't say there's really anything off the beaten path. It's fairly straightforwarded but I just wanted to highlight why we've put so much priority with a cash specifically as far as decentralized compute networks.
+
+00:50:00
+
+John-Alan Simmons: there's some other stuff that we kind of want to start working on as well. Kind of this idea of transparently creating micro VMs for users behind the scenes, when they Sign up for an application, the kind of be all and end, All of user-centricity and personal data repositories, is this idea that everyone's gonna manage their own servers? And, of course, you're not gonna expect, Little Granny to go out in some, Linux, Cisco, Sis admin stuff at 2 AM to make sure her box is running correctly and it's updated with the latest security patches, but the idea behind a cash which is basically, if you have some mechanic of pain for it, whoever, however, that is and you have a private key you can effectively just arbitrarily spin up VMs, which it's such a powerful function of what a cash provides and so you can kind of ultimately reach this Holy Grail of user-centric data and personal data repositories by leveraging both the Deaf or DV.
+
+John-Alan Simmons: Out of cash to build this whole ecosystem of micro VMs that transparently gets spun up for applications and for users as they kind of go through the sign of process without even them knowing that they have their own dedicated server sitting somewhere on the Internet. So that's some of the stuff I want to go through that. It is incredibly abridged and there is many things that I've lost over but I did want to kind of just seed you guys with kind of what other goals and requirements and functions of this database, how it relates to cash. And why we've kind of gone forward with this particular approach in this particular network.
+
+Greg Osuri: Yeah,…
+
+Tyler Wright: Excellent. Thank you. John Allen, Greg go ahead.
+
+Greg Osuri: thank you so much. This is amazing. I love database this for some background.
+
+John-Alan Simmons: You and…
+
+Greg Osuri: …
+
+John-Alan Simmons: me both Greg.
+
+Greg Osuri: I'm not an expert as much viewer and I love from a user standpoint that dealt with enough in my life too. to scale to a billion requests a day kind of thing. I also helped before for Eagle High,…
+
+John-Alan Simmons: Exactly, exactly.
+
+Greg Osuri: I help launch firebase, the first 10,…
+
+John-Alan Simmons: Really that I was born out…
+
+Greg Osuri: 15 years or so.
+
+John-Alan Simmons: Wow that's beautiful. Obviously Firebase inspired super bass and…
+
+Greg Osuri: So It reminds your demo.
+
+John-Alan Simmons: super based inspired Lee everything,…
+
+Greg Osuri: Really Yeah. Yeah. your demo reminds me of the early demo James.
+
+John-Alan Simmons: what's going on right now.
+
+Greg Osuri: The CEO fire says it.
+
+John-Alan Simmons: All…
+
+Greg Osuri: so, it has Right?
+
+John-Alan Simmons: bright future lead for us. That very beautiful. I'm glad to hear it.
+
+Greg Osuri: It has that sort of the feeling and I think what you're doing with a Christ's incredible. I love to. It's part of the road. Now we have this managed services where a user should be able to Schedule a container and…
+
+John-Alan Simmons: Yes.
+
+Greg Osuri: a database attached to the container just like…
+
+John-Alan Simmons: Yes.
+
+Greg Osuri: how you do on Heracles. That's the vision for cash, But do so on an orchestral way right into certain a very no without companies having to make deals,…
+
+John-Alan Simmons: Exactly.
+
+Greg Osuri: but rather protocols working with each other, So that's our vision and…
+
+John-Alan Simmons: Exactly.
+
+Greg Osuri: I think this fits right into that vision and database is the first obvious service, For fun application. That's managed And also, with IBC, you have interoperability, so you can have users pay in your own token or, now you have your SDC report, pretty much,…
+
+John-Alan Simmons: Exactly. Yep.
+
+Greg Osuri: right, so, it doesn't matter. So it makes it lots simpler in terms of payments.
+
+John-Alan Simmons: 100% very excited about that.
+
+Greg Osuri: And how your payments work, but you do whatever you can. If I want to stay, IVC payments, So, a lot of cool parallels. I can see how source can work with. The guy can really excited. I think we're gonna start sharing from this information online. that's okay with you. And thank you so much for the demo.
+
+John-Alan Simmons: Yes, absolutely. The more, the merrier, Thank you. Yeah, and regard kind of the manager service is
+
+Greg Osuri: Also. What's sorry what's your background? I mean databases. You don't just wake up and database, right
+
+John-Alan Simmons: So I just have this old adage that I don't know if I would preach to people once upon a time. And I basically said the Look if you end up that you have to build your own database you're doing something wrong and so that was years and years and years ago. And so now I've just eventually evolved into doing exactly what I said you should never do which is building my own database and scratch. Obviously I couldn't do this without the really great team we have here at source but foundation, I started in crypto and the mining days of 2011 2012 during the full blown Wild West s*** storm, that is crypto, apologies for my French. but foundation, it's like your science and number theory because I got into cryptography and then that just turned into building a alternative or competitive back in 2014 to Ipfs for a bunch of different reasons and then kind of leveraging that whole cryptographic structures and data models that I've learned and just leverage that into a database. And now we brought on this great team to make that happen.
+
+00:55:00
+
+Greg Osuri: The database, right? with the,…
+
+John-Alan Simmons: Yes, it is fully acid.
+
+Greg Osuri: with obviously,…
+
+Greg Osuri: obviously in a comments and…
+
+John-Alan Simmons: Yeah. Exactly.
+
+Greg Osuri: Try to adapt to the blockchain…
+
+John-Alan Simmons: Yes. Yeah.
+
+Greg Osuri: but it's cool.
+
+John-Alan Simmons: There's Alan internal mechanics and it's been another couple hours with everyone but I was standing.
+
+Greg Osuri: let's do, I have more tarot, Presentation allows the demo,…
+
+John-Alan Simmons: Definitely.
+
+Greg Osuri: but more power to a bigger audience, I think
+
+John-Alan Simmons: Yeah, yeah, I just want to add regarding kind of the managed services comment, it makes such a beautiful pairing to have a database like this, that is verifiable and freely cryptographically secure and all this great stuff when it comes to, a deployment in a managed services context in the cache because, it's so great to have all this, decentralized cloud, computing and open marketplace and all those great stuff. But then we build all these applications like anchor to blockchains. And then we attach a postgres server to all of it and the developers just in charge of the postcode server, we go so far that at the last moment,…
+
+Greg Osuri: Mmm.
+
+John-Alan Simmons: it's just a bit of a stumble and so knowing that you can kind of combine that whole ecosystem of decentralization blockchain that we work towards with off-chain data solution to actually kind of mimics those. Same mechanics that allow the user centricity and everything is really powerful for us.
+
+Greg Osuri: Yeah, but try to put a podcast. You end up being a blue cell. That's a disappointment now.
+
+John-Alan Simmons: Exactly. yes it's just a postgres database that happens to maybe talk to a blockchain but not really
+
+Greg Osuri: And that until some doing ridiculous poster in the entire state in a single validator. And it was just ridiculous. I mean obviously the homemade progress that approach, nothing personality feedback was it anyway,…
+
+John-Alan Simmons: Yeah, exactly.
+
+Greg Osuri: but I'm glad you're taking a very ground up approach and you're addressing some of the score issues fundamentally. I think that's was incredible. I don't really get a minute and but thank you so much.
+
+John-Alan Simmons: Yeah, that's perfect. No, thank you Gregory.
+
+Tyler Wright: Yeah. Thank…
+
+John-Alan Simmons: I really appreciate the comment feedback.
+
+Tyler Wright: Greg thank John Alan. I know we have a minute left. I'm sure there's going to be continued conversations. I'll work with the Source network team to see if we want to do a longer demo but it was great for them to introduce themselves and give us a high level of what they're doing. A much appreciated. Max had a question that chat, How do you authenticate to the database just want to see…
+
+John-Alan Simmons: Yes. Yep.
+
+Tyler Wright: if you can answer that real quick before we get out.
+
+John-Alan Simmons: Definitely. So the high level Tldr in that is that we have, one of the mechanics of our cosmos. APP Chain is to be this Global Authorization service that connects into the database and that's built on top of
+
+John-Alan Simmons: a data model called Relationship-based Access Control. It allows hyper granular levels of access control and it's one of the things that we're integrating literally this week so it's very excited to see. But basically means that you can have database or documents in the database encrypted you can have Such that individual peers can have access to individual records inside of a collection. The collections are inherently and automatically multi-tenant for that reason, but you can also specify the policy access and encryption of documents all the way down to individual fields. And that means you can specify that only these people have the access to these fields and those people have access to other fields and that becomes really important. for example a medical context where, your name and birthday might be not public.
+
+John-Alan Simmons: under the axis of the, administrative staff, but your Social Security numbers and blood types are only access to medical staff. And so that's based off a whole did system. It's built off of a vertical tree based X Control policy engine, that uses Graphtuples. And again if we had another couple hours but yeah it takes a long understand but you can get really rich into what the solo enables.
+
+Tyler Wright: Much appreciated again. to all the participated today. Thank you to the Source Network team for the demo. Obviously, Thank you to the Cloud. Most Console Team, Luna for all the work that they've been doing. And clients, I will certainly again, have more discussions about source network in a number, get an updates throughout the month of another sick. Clients meeting around this time next month, and obviously there's a number of other SIG and Working Group sessions, that will be happening. Anybody from North Source Network Team wants to have any plans sessions again? Shout out to NDS on this call, who is going to help make these notes and the transcripts and the recording available in the community. I'm sure we'll highlight this at the Steering Committee at the end of the month. But again, if there's anything that you all want to do in terms of meeting with certain groups over the next month, feel free to reach out to me and we can coordinate some of those conversations but much appreciate everybody taking the
+
+01:00:00
+
+Tyler Wright: Time today to have this sick clients monthly meeting was a great meeting. Anybody has any Thoughts comments about anything. This meeting. Feel free to throw those and discord and we'll just continue the conversation there until we meet next month. Again, appreciate everyone have a great day.
+
+John-Alan Simmons: thanks for the opportunity, Tyler
+
+Tyler Wright: Please our pleasure. Bye.
+
+Meeting ended after 01:00:29 👋
+