 Let me give a cheer. Let's turn my camera off for you guys here. Thank you all for joining us. My name is Brian Hobbs. I'm the CEO of Proven. And I myself as well as Demetrius Williams who's on here as well, my CTO of Proven. We are, as you know, reactivating the hyperledger meetup here in Atlanta. I think the last time this group met was 2017. So this is going to be interesting for us to get this restarted. We'd ask you guys to help as much as you can with this group because again, it's 688 people in this group. But as you can see tonight, we're very, we don't have as many people. So since 2017, we're not sure where everybody is, but we want to try to get them all back. But we want to make this experience fun, make it informational and definitely help people out who want to learn more about hyperledger. So today, we do have Arvin University. We have Matthew Russell from Arvin University, the RFID Lab. We'll be speaking today for our first group meeting here. We'll continue this and hopefully right now we'll do this on Thursdays at 6. We started at 6.30 today, but the intent is to be at 6 going forward every month. You still got to decide on the actual date, but we want you all as you did with the survey to give your input. So with that, I'm going to let maybe Demetrius chime in and then we'll hand it over to Matthew. Okay, well, I'm just going to be real brief because I want to get into the talk with Matthew given my opportunity. But as Brian mentioned, I am the CTO of approving and we're excited about getting this hyperledger group started again in Atlanta. And so we would love to get your help in making that happen. So please spread the word after this event, let everybody know how good it was and that, you know, get them to so that are encouraged them to visit us next month. So with that being said, I'm going to turn it over back to Brian and then Matt. I guess Brian doesn't want to say anything. I can go ahead and begin. Yeah, we'll let you go ahead, Matt. All right. Double duty here. So go ahead. Go ahead, Matt. All right. Thank you. Cool. So I'll go ahead and give a background before I before I start presenting. My name is Matthew Russell. I've been currently a student at Auburn University. I work for the Auburn University RFID lab. And I've worked there for about two and a half years at this point. And what we focus on is primarily identification technologies and how we can use those from a technical perspective, but also what the business value is for using those in the retail and apparel supply chain. But then we're also looking at other spaces to like food and grocery, as well as aviation. So being the RFID lab, we focus a lot on RFID technology and how we can use that kind of in different use cases. But the projects I've been on, I currently lead the data initiatives at the lab for a lot of our retail and apparel projects. And I've been kind of on various projects throughout the years, but really focusing on that and focusing on the data capture piece, the serialization of items, the data quality, and then also the data exchange. And that's the data exchange piece is really where we see blockchain and hyperledger fabric specifically kind of as an opportunity to be used. And we ran a proof of concept to actually evaluate that in the supply chain. So I'm going to go ahead and share my screen here. It looks like I need sharing privileges. So, I guess, Brian, if you're able to give me screen sharing privileges, that'd be awesome. Okay, should have it now. There we go. All right. And I know we have a pretty small group here. So if you want to stop me at any point as I'm going through this and ask questions or you want me to clarify anything then definitely feel free. It's a relatively short presentation, so it's not like I'm really trying to get through all of it. But all right, we'll go ahead and jump on in. So, like I said, I'm with the Auburn RFID lab and we really, you know, have grown through the years looking at how we can use RFID in the retail and apparel space. And over and over again, we'd work on these projects where we're looking at kind of the entire supply chain. So all the way up to the factory and then down to the store and then all the distribution centers and warehouses and everything in between. And we'll be trying to capture these RFID tags that are being put on clothing items because kind of throughout the years and a lot of big retailers have pushed it. There's been a lot more demand and kind of mandating of putting an RFID tag onto a clothing item at source. And they do that for a couple of reasons, but primarily in the store, they will use it for what's called cycle counting. So if I'm, you know, in a retail store, like, you know, I'll give Walmart as an example. And I'm in the apparel section. If I want to get an idea of what I actually have on hand, then traditionally I would go through and scan the barcode for every item that's in that store. And none of us want to do that. No one wants to do that. That takes a long time. And they only do that a couple of times per year, if that, because it really is a pretty involved process. And so RFID, what that does is allows them to not only assign a unique identity, but also read those tags in a much easier fashion. So, instead of establishing a line of sight with every item, then they can, you know, put out a radio signal with either a handheld device or a fixed infrastructure or something like that, and then read all those hundreds of tags in a matter of seconds rather than just a few tags. So then they can get a better idea of specifically what they have on hand more often and drive up their inventory accuracy numbers. So that's kind of how RFID began in the apparel space. And as we began looking farther upstream, we realized, well, you know, we're, we're tagging these items at the factory, and they're going all the way through the supply chain down to the store to be used. But why can't we capture this data at every point along the way in the supply chain, and then leverage that. We ran into a problem of, there's not really a great way to share this data or change this data in sort of a democratic way. When we begin looking at, you know, complex trade networks, you can't just say, okay, we're going to use this one platform or have this one solution provider. So we needed something that's that's sort of distributed and easier for individuals to tap into. So, all I guess to say that we put together a working group to kind of work through some of these problems, and then jumped into an actual proof of concept for what that would look like and you can see here we had Nike, Macy's, Kohl's, Herman K, and PDH as the partners that were on that project. And we wrapped that up in December of last year. And then now we've moved into the pilot piece, which is what we're working on and I'll get to some of the details on that later. Kind of like I was saying earlier, what we're traditionally looking at in the supply chain is very siloed. So this ASN here at the top that represents the traditional flow of data in the supply chain, generally, and that's really point to point so a manufacturer would say to a supplier, Hey, you know, here's the advanced shipping notice for what I'm supposed to be sending you. And then the suppliers pretty much blind until the products actually show up. And then even then it's difficult for them to validate and double check everything to make sure it matches up so they really just trust the manufacturer and what they say they're sending. And the same goes from a supplier to a retailer, or even a retailer to their own retail store. And then shifted from that's, that's called G 10 data. That's a global trade identification number. There's a lot of acronyms on the screen. So, you know, bear with me here. But all that is just like a barcode again, that's a quantity level basis. So the manufacturer would say hey I'm sending you five white t shirts. And that's the only, you know, that's as granular as that data gets. So, like I said we've kind of shifted towards the RFID data, which is serialized. So then, okay, you know we can distinguish each item, it has an identity, every products I guess. But even if we're capturing that data for each of these unique items at every point, we're still not sharing it it's being siloed, which you can see here on the bottom of the screen. So we're not leveraging it to, to really communicate what's going back and forth. So, when we begin to look at the blockchain and how we can implement that this is kind of the general idea of, rather than having the serialized data be siloed at each point. We create something that's no longer unidirectional, but allows greater visibility for the supply chain partners. So the data to the blockchain network and the manufacturer can say, Hey, here's the individual items that I'm sending you the supplier and supplier can see that before they even receive them. And it is so much easier for a supplier or retailer or anyone to validate and say, Okay, let me make sure I got what I received when they have a known ledger of what they're already expecting. Because then they're not operating blind. So that's the big idea. And there's a couple of different business values or reasons that you know individuals and supply chain might want to do this. And that is these numbers here. And so this is from the retail and apparel and I think it includes grocery as well. But the supply chains, and we're looking at about $100 billion of counterfeiting cost each year. And, you know, this is pretty ambiguous. A lot of this is also unknown claims is estimated to be around $35 billion and shrink at about $50 billion. So claims is when a supplier would send a retailer or any manufacturer supplier any two partners would have a disagreement on what's being sent. You know, so retailer would say, Hey, you know, I was supposed to receive 10 but I only actually received nine items. Let me let me issue a claim for that. And then shrink is really lost or unaccounted for inventory. And that would be due to theft or administrative errors. I think the idea is that we really have no idea. Otherwise it would be classified differently. And so these are huge problems in our supply chain, and we don't have great visibility into, you know, how we can resolve these or what's actually going on. And that would definitely ease some of the pains and that's what we believe in what we feel blockchain can help us work towards. So these are the partners that helped or contributed data to the proof concept. We split them up into three different data streams or three different channels by that. So hyper ledger fabric offers the channel function, which allows us to kind of create a partitioned off space where Macy's for example wouldn't be able to see calls data. So it was vertical going from encoding data at the manufacturing level all the way down to the store. Herman K Macy's were a partner pair and then pvh and coals were also a partner pair. And the primary focus of this is really looking at that handoff point between the two partners. So a brand and a wholesale retailer, because that's what we're most concerned with that's where those claims are going to arise and those disagreements come up. So when when Herman K for example, would send something to Macy's and Macy says hey wait, you didn't send me what I was asking for, and then they just go back and forth on email trying to figure that out. So that's kind of what we were looking for and you know how we could ease those those pains with blockchain. So we broke it down into three different steps. The first being to identify the systems and the stakeholders. You know, looking at the supply chains of each of the partners, looking at who controls that data. Who, you know what points can we even get data where you capturing data in the supply chain, and what sort of system systems do you have and how can we gain access to that. And once we identify those data streams of, you know, where are you reading the RFID tags, or the case codes, and different things like that. We really had to work to standardize them. And that, that's a huge part of this and that's what we spent a lot of the time on was, you know, we have to make sure we're speaking the same language if we're going to have a common network among us. And so we use EPC is, and that's an event based standard that's been defined by GS one, and that sort of captures the context of what's happening to a product in the supply chain. And that was was hugely important. And that's what we feel is going to be the future kind of language of business at least in the supply chain for product based events. And then the last step was to actually integrate this data. So we've identified where the data is coming from how to get it, we standardized it, and then then we actually integrate it into a common network. And that's a hyper religious fabric based solution that we put together. So again, this is a focusing on EPC is here, it answers the questions, what, when, where and why. So what item is going from point to point, at what time, what's the location of that and then why, what's going on is this item being shipped or is it being received. There's a whole kind of vocabulary that GS one is defined. And I know that standards are usually boring, but they're pretty necessary. And so it's definitely something that's, that's really important. When we begin to kind of move towards a common platform like this. And then finally, we integrated into the blockchain. So this is kind of the exciting part, the new part. And we worked with the IBM blockchain platform. It's relatively easy to use, we didn't have to code a lot of the parameters of the network. But we did have to put together client applications in the front facing web applications actually, both right and read the data that was going through. So, each partner that was contributing data to the project could go in and upload their own data, but then also read the data that's in their specific channel. As soon as their partner that they're working with or uploading data with put something in and they can say, hey, you know, I can make sure that the items that I sent have actually been received and can see the history or see the story for every single product. And that's where we really want to get to is a common place where a supply chain stakeholder can go in and really see the entire story of each of their products. They're going through the supply chain because right now they don't have great visibility into that. There's a lot of data here I can send out this deck afterwards everyone that was on the call, or not a lot of data but a lot of information, because I think it might be helpful if you've got more time. And then finally, ultimately we had about 223,000 RFID tagged items that were accounted for from source to store and that's, and that's kind of at various points, not all of them were all the way from manufacturing to the retail store. So that was 223,000 items that we saw in our blockchain. So key takeaways. We were able to see that, you know, blockchain actually does work, we can use it to exchange data. But the reality is that the transaction speeds are pretty low. What we were able to get to was like 25 to 30 transactions per second, which, you know, I know now with fabric there's been a lot of testing for higher transaction speeds and we obviously are a group of students from a university. But we did work with the IBM team and some other individuals on this and compared to traditional systems and it is by default. It's efficient. And so I think it does need to be proven as scalable before we begin actually using it to exchange data. Because that's what we were doing was taking data itself and exchanging that, rather than using it as like a validation of some sorts and only storing hashes or something like that and that's what I know a lot of other projects are doing. And so that will probably be the route that we go in the future but that's something that we're still kind of exploring. But then beyond that, EPC is key, or really any standardization is key. I said it earlier but I'll say it again that a common platform requires a common language. So, you know, regardless of what sort of network you have for those data inputs you want to make sure that those are in some sort of standardized format, because that makes things a whole lot easier. So data quality is paramount. And, you know, as they say garbage in garbage out throughout this entire proof of concept that we ran we didn't focus a whole lot on that quality of the data that was actually going through some of the some of the RFID capture systems that were reading items were reading 70% of the items, and they only had them on, you know, a small subset of conveyor lines in those distribution centers. And then we when we really look at, you know, what does this look like in production, we really want to make sure that that data we're sharing on the blockchain is is accurate. And that really encompasses everything that's that's going through. Otherwise, if we're sharing something that can't be deleted, and then we have to append, you know, we don't want to have 70% of our only 70% of our data being correct. So that's what kind of going out of. I've got the launch the chip pilot is kind of what's next, going into the chip pilot we're really focusing on that data quality piece because that that's really important to get the business value out of an exchange mechanism like the blockchain. And you can see here on the right side of the screen. Again, we want to look at the scalability and the performance of blockchain. I think we're still a little bit early on for for some of the demands that we're seeing in the supply chain space, although there definitely are use cases that the blockchain could be applicable for. And also look at vendor neutral networks and multi cloud deployments, because when you know when we're in a space where there's a lot of players, and someone has to be controlling that network. And that's going to have to be a neutral sort of entity, because, you know, no one, no one's going to participate if you say hey, you know you have to give all your data over or you have to control, or be controlled by this one organization, that they may not trust. And that's a really big inhibitor. And so a lot of people have brought up that they're concerned about. So, you know, making sure that we really do have a neutral environment that can support different cloud systems or different cloud deployments. And that was a huge concern or I guess, key point from a lot of our participants in the project in the workgroup that we've had. The next next for us is also to look at the pilot, which is really focusing on, you know, now we've we've identified what it looks like to exchange data, but if we were to exchange that. What are the business implications of it. And how can we reduce those things like claims and shrink. And what are some of the benefits we see from capturing serialized data and communicating that in the supply chain. And that's really what we're focusing on now, but we found that we really need to get that quality data, and make sure we're capturing everything effectively, before we begin focusing on how we can actually share it, or how we're going to share it. And that's, that's kind of where we've been the past few months on this project at the lab. And that's obviously been slowed down with with COVID and everything so we're going to be continuing with that into the next year. And that that's really adding in that last step here of, you know, identifying standardizing and integrating but actually analyzing. And that's the end goal is, you know, if we're using blockchain, taking things like smart contracts to leverage the data that's on that chain, and actually analyze it to derive a business value for what does this mean how does this help us, because that's what's going to get a lot of the supply chain stakeholders on board. You know, that's all they're looking for is, Hey, how much money is this going to save me, or how can this make my life easier. So that's the part that we're really looking into and focusing on now. And here's some more information about that smart contract piece and how that would actually work. This is pretty complex, but essentially, you know, taking the data on the chain and verifying it with, you know, what's supposed to be sent versus what's going to be sent and then creating an alert for if those don't match up. But again, I can I can send this out, if you would be interested to take a closer look at it. So that's, that's all I've got. I know I kind of rushed through that so if anyone has any questions, or if I didn't clarify anything or you want to go back, feel free to speak up. Please don't be shy. Definitely want this to be an active interactive experience and thank you Matt Matt for for this great presentation so hopefully we'll get a lot of questions here. I have a question. Yep. Okay. How many different distribution centers were there. Oh, I have to think back. I think we had. I think it was me, let me run back to the slide with the partners and try to recall usually have a slide that the kind of details the individual ones here. So we would have had one for each of the participants so five so five. Okay. Yeah. And when I guess, having worked some of the IBM in the past I was curious. Well first I guess how many nodes did you end up having. That's also a good question in the actual network itself. Yeah, or more specifically, I guess, well, yeah, did. I mean, we had a node for for each point that was contributing data into the chain, which thing would have been. I have to go back and look, it was either two or three for for each partner pair someone had one. And then that was again split out into those channels. But I had to look back and get back to you know but it was multiple per partner. Okay so you maybe had as many as 10 to 15. Okay, and I mean I was asking that question for a number of reasons but you said you worked with IBM on this did did go through all IBM type nodes was provided from them or did you other use other things outside of IBM other nodes. Yeah so they were all provided by IBM, and that was through the IBM blockchain platform, and then it was supported by the IBM cloud. So that's something that, you know, we kind of got a little bit of pushback on was was that truly blockchain if it's all, you know, kind of centralized in that one space. And so that's what, you know, looking at in the future, how can we bring in other solution providers or service providers in different cloud deployments. I think that was, that was a big concern, or not concern but just kind of something that the end users were asking for. Yeah that's kind of why I was asking the question you know because sometimes they, they say they're giving you nodes but they put you in one particular farm, if you will, and not that that's necessarily bad it's just that's been our experience. That would might make provide some indication of you said you had a low number of transactions per second. So, you know it's, there's a lot of variables there but that's what I was asking that question for okay. Thank you. Yeah. Yeah, and that's, that's we didn't. Yeah, go ahead Daniel. Oh, can you hear me first I'll just want to be sure you can hear me. Yes. Okay cool. The question was pulling back to talking about your web app and the design of that and stuff was, how did you go about designing that and everything. Because, I mean, most of the blockchain tools like composer from the research I've done there, it was deprecated back and I think what August 2019 I think it was. And did y'all just build it from the ground up or did IBM help you out with that or like and how'd you kind of link it through with the your blockchain client in the platform and everything. Yeah, so we, that is something that we did build from kind of the ground up independently. And I wish I should have brought some of our technical team members on to here because I know they definitely be able to answer some of these and a lot more detail. So we, they were able to communicate so we would upload to the web app. And that would go through to the client and we kind of tailor each client to feed into the blockchain itself for each different data stream that we were getting so that would be like each individual node. So we had smart contracts to upload the data in the format that we kind of predefined. But we did, we did build the the web app and the client independently. And I don't know, I believe they use API is to communicate with each other. But I'm not 100% sure. So if you want I can get with them and get a little more detail. Yeah, no that'd be great. Yeah, because, like, from my experience, like, they had a great tool there, I don't know if you ever played around with hyper ledger composer, but it was a fantastic tool. I mean, you could build a whole blockchain application I mean, as bare bones as it was I mean you could build in 45 minutes, and then like they just got rid of it and it kind of really from a front end perspective of developing a web app. Like, there's nothing that has come out that I am aware of. There's nothing that's come out since that I mean I've heard of many fabric which is like I've kind of heard they kind of dabble on that but it's not like a tool that is dedicated to helping you develop a full front end web app essentially to interact with your blockchain platform. I'm just curious. Yeah, yeah and ours was ours was pretty crude. I mean it was nothing, nothing fancy but it was essentially, you know, quick here to upload a file, and then click here to get an output of the data that's on that but the their IBM's blockchain platform itself is in a way it could be kind of the front end, because it allows you to see a lot of the stuff that's actually in the network. But yeah, we haven't looked really into some of the options for that either. So your interface was real basic like you said it was like, like an upload a PDF for an example and you just click and then upload it and then I mean there's probably a little more to that obviously but the gist of it was like, in the transactions kind of what the history of the distributed ledger what have you and then like upload a PDF or something to that extent. Yeah, yeah, and it was they the partners would use an API to tie into that. In some cases, to actually kind of automate that process. And that's, that's really what pushed us to, to have that web app was, you know that makes that process a lot easier. But I know that was we, we ran into the problem of like querying a lot of different items at once and ended up having to. I think at the time like mirroring what was on that chain. You know, letting us that function, because we, we really honestly we didn't have the time to really build out and explore that function as much as we would have liked. Right. Okay, cool. Have a question similar follow up. You know what application development tools they they used on the front end I mean did they use JavaScript or any particular development tools. Um, I'm not certain. I feel like it was but I'm going to take note of that and I'll confirm that and then I can, I can send you an email with the information on it. And the database that you used. Did you, most of these are non sequel, did you guys use a non sequel that so what was it. Um, I know it was non sequel. I don't recall. Well, I think by default hyper load or fabric, I think uses level I think is what it's called and then you can use couch I know as well. Yeah, so I, it was, it was couch. That's what it was. Okay, thank you. Yeah, and I apologize. I'm not the most technical one on our team so I'm, you know, I'm not able to answer a lot of these questions but I can, I can get back with my team and answer anything that they have, and get back with you guys. Hey Matthew, I have a question for you. Yep. Yeah, if you don't have the answers you can always get back to us. You use it as a nice presentation so I have just six months of experience and I'm solving something similar problem in supply chain electronics. So, you bring up a very valid point. The carbide is something coming out without the data is a malicious if it is counterfeit data or recycled product data. We are not going anywhere. So, in this case, I see the slide the brand or the retailers of it's the data. But let's say most of the companies they are using their outsourcing their manufacturing. They don't have any specific country here but it's a reality they're outsourcing so they're outsourcing partner sometimes it's their entity in that country or sometimes it's a totally different country they just put a label, and they just done. So how to make sure that if manufacturing entities outsource what the data they are providing is legitimate. So, yeah, make sure in this blockchain. Yeah, and that's, that's kind of a difficult question that we've run to because some brands may be using hundreds of different factories to actually produce their items. And it's very difficult to manage the data that's coming out of those especially when it's going on to a chain because you know you can say, Well, once that data is on the chain and it's secure and we can trust it, but you have to trust the quality of it before it even goes in in the first place. And so that's, that's a difficult question that we're working through right now. And then we're beginning to look at is, you know, how do we honestly you need to have some sort of auditing program to go through and check some of the data that is coming out of these locations that are all over the world. And how do we get to the point where we can actually trust it. And you know that kind of probably varies or depends based on what sort of data that you've got coming out and going in working with RFID data. We're working on kind of a general process or a way that you would go about actually auditing the data and verifying that it is, you know, legitimate and is accounting for everything. But, you know, needless to say that it is really important to think about, because I think a lot of people hark on the security aspect of blockchain. But, you know, that doesn't help change malicious data before it even goes in or incomplete data before it even goes in. So I got one quick question. It's kind of a crazy question, but I figured I'd ask it anyway. So why did you pick hyperledger fabric versus something like hyperledger grid because grid was from what I understand was designed for supply chain. When it was being designed so I was just curious why fabric versus like the other platforms or frameworks. Yeah, so I, I wasn't even at the lab when this during the kind of inception of this project. And so that was, that was before my time when they actually decided hey we're going to roll with hyperledger fabric. But that was also before grid was was really released. And so I think fabric at the time was kind of the best option for a private information blockchain. I think it just made the most sense to the earlier that was on the project at the time. Right. I mean I know fabrics probably the most supported I guess would be the right word of all the frameworks out there. I know it's all to picking up some recognition to send was just curious. Yep. There's a lot of you know when you look at the GS one standards for barcodes there's a lot of there's about 11 I think primary barcodes that although there's plenty there's in the 10 or 11 that's widely used. Do you get the different people in this group to decide on a common barcode or did they not have common barcodes and you just you guys just write a application that would interpret the data, you know basically a filter mechanism. Yeah, so we, it was a mixture of both. I think it was a mixture of having conversations of, you know, we, we want to get to a point down the road where we have this standard. And right now, particularly with the serialized data in the RFID, you know everyone's kind of doing their own thing and exploring what that looks like and so we were able to push some of them to adopt that GS one standard the EPC is, but it was about half and half of some of them chose to adopt it and use it and send that through. But the other half, we just ran it through a an application as part of the web application was it would actually convert that data into the proper format before sending it through to the client. And that was that was another function of why we wanted to have it go through a web app in the first place. But ideally, you know long term, they would produce it from the output. It would be good if it was really standardized but I figured you probably wrote some kind of a conversion interpreter, or something like that. Now, I mean I know it's the university that did that but who actually paid for the servers and stuff IBM I mean they're not. They're not cheap usually after they get you out of the sandbox I mean, did the university pay for that or did they have a contract or, and then how much do you know how much it may have cost. Now, yeah, so that was like a research education grant from IBM. So the university did not have to direct the pay for it. I don't recall how much it actually costs. I would have to reach out to them and get those numbers. That's something that we didn't really look into. But you know I've, I've heard from people that some of the prices that you know IBM Oracle or some of the other providers, some of the prices for these services are kind of up there. And, you know, I think that's something that we could have taken into account when looking at how does this compare to traditional data storage or communication methods. But yeah, I don't we don't have a an exact amount. That's a really good question now. I think that's a valid point. Thank you. I think I haven't looked very recently but last time I looked I think for the production sizes 1500 a month but don't hold me to that and then there was a smaller tier that I think was 500 a month I think don't hold me to that but that's kind of the ballpark I'd say about what they're charging you monthly. Yeah, and that's what, like, kind of moving to the second phase of once we kind of get down the data quality part, and begin to look at what is the actual business value here, like how much can we actually reduce these costs is when we were going to really jump into. Okay, well, you know how worth it is this, because that's a big question that we're getting from all the end users is, you know, yes, this is, this is all right, but how is this going to save me money and what does this actually cost to do and how do I do it. And, and we don't have great answers for any of those now, just because that hasn't really been explored by anyone yet. Any other questions. There's no more I've got some of these noted down so I can get with my development team and then I think I should be able to get your emails and send you kind of a follow up from this. Well, that would be great. Sure we have somebody made a comment in the, in the chat. Or a day long. Say multi 12 maybe not quite sure. I think you might have said that when I was talking about the grid between grip grid and fabric, maybe I think as well as all that pop up. Okay. Okay, any more questions anyone else. I got one other question you said you on your, on your couch. You're pretty sure they use couch. Did you have some students that were really well burst in couch, or, you know, how did they start using that, or did you use somebody outside the university to develop the database side. So that was, that was students on our team. I think the one that primarily managed that actually graduated so I have to reach back out to him, specifically but I, from my understanding they had pretty significant database management experience, just through classes and other projects and things like that. But also a lot of particularly with fabric, you know, they had to, they had to learn just about everything because it was pretty unfamiliar. And so, I don't know exactly how much they, they kind of had to learn on the fly. But I know that they did have experience using kind of database functions like that. Okay, thank you. Any other questions. All right. Well, well, first of all, I would like to thank Matt again for the excellent presentation that he gave tonight. Definitely was a lot. You know, good information here and I hope we all can take some things back from from what he presented and maybe do some more investigation. He's going to provide the slides to the group. So, and he's got a couple of follow up questions that he's going to research and get back to you guys on. We'd like to, we're going to probably send out a survey just to get some feedback from you guys and how we can approve and what you liked and what can we do better with but we again we still encourage you to reach out to others who are interested in this space, because we're really trying to grow it again here in Atlanta and support the community. So, if you want to look out, we're going to do it again next month again at six o'clock and not 630. And you'll get an invite for that. So, with that, you guys have a great evening and thanks again for attending. And thanks again, Matt, really do appreciate you. Thank you, Matt. Thank you. Thank you all.