Speaker 1 0:00
Can we just ask either Harry or mark for the viewers at home again, you guys are an IIoT platform. What are you here to prove? What was the goal? What did you build before
Speaker 2 0:08
Eric bolt that question on the right, talking more about that platform. Okay, transitioning. Because one thing that's very important stress we are very strong on the factory floor and on the operations many IoT platforms, the background of the company and the design your IoT guys are cloud infrastructure guys. They start to go to the factory floor. But factor four is a real interesting place. Absolutely, I don't want to show is another word in the live broadcast. We were born doing factory floor startups really the hardship, and we grow to be an enterprise outside solution also. So we do have many things on the system that are very true requirements for industrial solutions and for factory floor, okay? That cheaply, someone, that's only IoT, will not do that. I will always say that you are also a scalar platform, and admitting an i IoT, okay, that additional eye for industrial is very important to our federal positioning. Gotcha.
Unknown Speaker 1:20
Now about provid
Speaker 3 1:22
showcase, yeah, so we really wanted to show off the platform's capability, yep, how a manufacturing company could use it effectively to deal and solve some of their challenges and issues. As the whole premise of prove it is about here. Our solution is really starts with the connecting clock and showing kind of the ease and elegance of the way we connect to this type of environment with MTT, SQL and OBC. So start with Connect, collect, move to transformation data. We have an internal uns that basically can become aware of external unses, Utah, internally, we could transform the data through scripting. We were just modeling of all sorts. So we showed off some of that platform, and then the result is the visualization. So the deep kind of visualization and interactive dashboards that would help this specific type of manufacturer make real time decisions while they're producing so that they can effectively improve their efficiency and quality.
Speaker 4 2:50
Gotcha. And so ultimately, for the viewers, I don't want to make the assumption that everyone understood the assignment. Approve it. You were given a data set via a uns, which is ultimately MPTC broker in this case, and Mark maybe to expand a little bit on what Harry has mentioned, could you explain to us what you have built, and maybe, what kind of tools within, like tat soft, have you had
Speaker 2 3:10
on the concept connect, transform and present a point of the day on the requirements we put to this solution? What going to prove here? Yes. Number one, we have 36 or 39 companies publishing data, the connection you can do the traditional Scalise that you map the variables need to be able to do what we do here, that who have the size to put new assets here, our application recognize automatically and to be able to use that data without having previous knowledge that data will be there if you did application, to be able to show the stats the OE for production lines, effort back machines for eliminators, whoever publish that type of assets, the application should be self aware and dynamically responsive on the connectivity. So don't need to stress anything, because that's what went to prove on the first parts, on the connectivity, on the transformation. It's also important, because even you don't go to analytics, let's stay even more simple, some comps are published always zero to one. Some comps are published zero to 100% some companies put the line in the machine, not on the 95 standard way down the tree. So we had to do lots of and that's a real life application, exactly sure I thought it really happened. Yes, we
Speaker 4 4:42
talked a little bit about like before the stream, and I really resonate with that, right? Because you get to the plant floor, even the same machine, OEM is not going to have the same exact data between different machines. I think transformation is really important here to
Speaker 2 4:55
find besides Python that many people use, we have the full.net Sharp and vb.net, very powerful and high performance client side and server side codes to manage that. Then on the presentation layer, what were the requirements to our specification? First of all, things like switching quality schemas. Comma this year that should be out of the box and should be able to apply our application. You can call these schemas without hand, do any programming for that. Beyond of that, the performance of the UI, whatever running locally or in a URL from a mobile device, because you came from factory floor. We know that 123, settles. Is not acceptable. When you have an operation machine, the guy clicked a button and he does not see anything happen. How much what you say by experience? What expectation of reaction,
Speaker 4 5:58
usually over 300 to maybe 400 milliseconds. It becomes pain. It becomes like seconds. Yes,
Speaker 2 6:04
so I don't know if the camera will be able to pick up properly, but this same application, when you open the mobile right URL, you still have the same performance on the web page. So we are able to deploy very rich graphics, either mainly or through web or different devices, and adapting properly that UI to the specific devices. So if you want the desktop, you have all the areas in front of mobile, you have to reorganize the data, and we do that automatically also, so the same solution can be deployed in manufacture. So both were the details. Oh, why those three requirements? They are little more complex. Because I say, Yeah, collect or solve and present. In some sense, everyone just throws in the head.
Speaker 3 6:59
I thought to make one more comment too. Is the entire, our entire application is done with our platform. Okay, we're not using any other third party tools. We're certainly, we're connecting into the data, but often full application is done with us. Now that in the real world, it's not always just us. So we're that open, flexible part we can fit in the most manufacturing or industrial landscapes then. But for Pruvit, we were able to do everything within our own Bucha
Speaker 2 7:34
and in Mark and that we understand as much as possible. If your tool is important, we should have that built in. I will give you specific examples. We have 70 communication protocols that we install with the products. We have a built in MTG broker to help your tests. You have a built in OPC server. We have a built in OPC publish simulator and NK simulator. So we have lots of tools in all areas. So the connectivity will be to really to publish data to another application or collect data. But we cheaply don't need to go as a tool, see, to feel some places that, okay, cannot do that. Let's pick up. Let's become not to know cheaply platform even the dots net, we have the actual and the compilers beauty.
Speaker 4 8:27
I want to maybe ask Harry on that side, right? Because I think of the with so much flexibility also comes maybe the decisiveness of what you can use this for, right? And ultimately, I'm curious, when you're having conversations with some of your customers. What are they looking and seeing taut soft as is it to build their HMI systems? Is it to build their SCADA systems? Is it to build their mes? Is it as an analytics tool, right? Or is it all of the above, maybe? And what is a general use case? And our
Speaker 3 8:56
number one challenge is it's all of the above? Yes. So
Speaker 1 9:01
I think that's the best and the worst thing about a platform, right? And we've got several platforms here, and that's what I think about all of them, right? Like it's the best and the worst that it's a platform, right? The thing that you start building your HMIs on is probably not going to be the thing that you think of as your IoT platform, right? Like it could do all of it. But many groups think about the tool they may own as one particular thing. We can
Speaker 2 9:25
give some general guidelines, yep, where we really excel. The more you have real time data, okay, we are great, the more closer you are to the real fact for data, yes, we are better, the more you are unwilling to compromise your specifications. You do an application around the two. That's the old school, the newest who you are. Forget the two. I will achieve those results. The two is not the main actor you should be able to deliver. Those are some of the. Cheaters were likely. We've been and also very scalable. Okay, you have applications running like Cisco filters using only 500 megabytes of memory, and distribute applications with half million points in electrical distribution. So in general, if your solution has some of those requirements. Yep, it's more likely. And one wanted bots. That's very important. We support the full dots net to Windows with active direct integration and the.net H deploying any device in minutes in Docker, whatever. And bean.net is very important because it's easier to connect, to have better security managements. You can more powerful programming languages. That's another I can true
Speaker 4 10:48
answer. I want to go back maybe to some of the important points that you mentioned and dig into just a little bit deeper, right? So you mentioned connectivity, and I think that it's important to recognize your prove it you're going to uns broker, which I would argue is not the common connector, at least some of the clients I'm talking to. So maybe, what does the typical deployment for you look like? And ultimately, what are you pulling data out of? And what can you talk to? Right both, I want to say, on the plant floor side, but also maybe at the ERP level or other systems. Yeah,
Speaker 3 11:20
I would say typically, we're connecting directly to some kind of control, or IO data collection device at the north side. I guess we'll call it. And then, in today's world, a lot more requirements to be able to take that information and publish it, yep, to either more open like MQ, TT based system, or specifically to Microsoft CRP or SAP and those connections. Again, we're both been on the southbound and also so we can do we have several ways to do it, but RESTful APIs all sorts of ways to communicate openly to a variety of systems and
Unknown Speaker 12:08
only to make shift down
Speaker 2 12:12
in the last few decades, it create more than 250 communication protocols. So one thing that you are very low also whatever protocol need to have an inducer plans. We were able to bring that back for data into our platform, and then we go up with all those many options, Harry pointed out, but whatever is counter logics, module, busses, VP, legacy, Bailey season, we have all those industrial protocols built in to do that. First level, station, real life.
Speaker 3 12:50
COVID seems to be a really popular thing lately for our interest prospects and customers, and we built a very nice, discoverable interface to the code assist to the LC, interesting
Speaker 2 13:05
people. That's not from any code sees. They have already 17 brands of PLCs and controllers absolutely based in their engine. I
Speaker 1 13:15
want to talk about that a little bit more about what you guys are seeing, what the demand in the community is. So here we're at Pruvit. We're talking a lot about MQ, TT with the UNs central broker. Are you guys getting a lot of demand, a lot of questions about that? What sort of communication protocols, what sort of conversations are you having and have the has that changed over the last couple of years
Speaker 2 13:38
about MTT is, at least our experience not yet a reality factory floor, okay, but in many large, very conservative customers, it's become a reality. On the high end, yes. So we do have many applications that are running edge devices, etching data from control logics rock with devices, yes, and publishing up to a broker, yes, catching data from Universe Universal Robots, absolutely. So we start to have lots of conversation to be able to have that connection up to some sort of MP chip protocol. That's really start to happen at the at that level,
Speaker 3 14:21
yeah, I would say a good 50% is still a traditional, we're talking to a BLC, absolutely providing SCADA, like application. Now more of a requirement to say. Now, where does that data go? Absolutely there, for sure. But now the other 50% is really this, I will call it newer, more modern architectures. You know,
Speaker 1 14:45
absolutely I would agree. I see more and more groups having MQ TT in their specifications. Or the conversation about MQTT happening lots of times we look at those legacy controllers and we can't natively publish, uh, MQTT. So typically it's some sort of OPC UA server, and we're making that connection. But then we're also looking at perhaps smart sensors, and can we go publish directly to a broker? But I certainly see more and more demand, and I imagine that we will continue to see more demand, or at least conversations around MQTT, which I think is very good for the community as a whole,
Speaker 3 15:20
yeah. If I can make one more comment, there is, I was listening to Arlen yesterday and that he comes out of the oil and gas industry, yep. And we see some pretty good in that marketplace. People upgrading their status systems, yeah, their infrastructure around MTG, absolutely. Yeah. We just recently did a fairly large project with one of our integrator partners running this was a Cisco router out in the field doing the connection to the PLC classic back to the T so that modernized their whole communications architecture future. And it was, it was a there was fun applications that view. Mark was very involved with that.
Speaker 2 16:04
And it's important to explain about architectures. People won't be over excited about nkchi Chip, because if you think about our edge device publishing and pitch data up like the solution described, there is no reason at all to try to replace the protocol from the ad devices to the fields by mpg, no reason at all. In fact, in many projects, does not even make sense. That's the difference between I IoT in booster IG or generic IoT. I'll explain briefly, if you have IO Cha applications work with meters, house meters, going to the application, you can very easily add and remove meters, yes, in the system, and you are not going to create alarms if a meter go down because that house are sold, they're just stopping collect that information. If it's a more traditional production line farm oil and gas, when you are talking about the sensors, you don't want add or remove sensors dynamically. Wow. In flat manufacturing, go through fg certification, yep. So the reality in the factory floor, most of the industrial process. We don't want to do things that dynamic. That's one of the many benefits on mkg chair at that level, if it's industrial plants. But when you have that your edge and start do data consolidation, data aggregation, the enterprise level, all now mkg is great, so don't try to replace I saw something. Oh, I I have this, just sensor, this meter for temperature that you publish right away. Mkg, Chee, that's not really the main focus, right? You can do that for fun, but that's not really the point that MK Q was created for definitely
Speaker 4 18:02
agree. Let me ask you, maybe on a more fundamental level, right? If I'm looking to create maybe a proof of concept, like putting myself in the shoes of a customer, right, I'm a senior engineer in a manufacturing facility. We probably don't have a modern SCADA or mes. I'm looking to create a proof of concept, right? Can I go to tatsoft and maybe create some kind of a demo for my upper management to sell them on this idea, and maybe what does that look like? What do I need? Is that just the software running on my laptop, or is that a server no roll out? What does that
Speaker 2 18:33
look just the software on your laptop and you can download from websites the full configuration with no requirements for license for the population too. And what you can accomplish, probably should show what is for per week. Sure this application we created and around for one or two weeks, the main development was around one week. Then we keep publishing hand there, but it was in two weeks. So what we did, in this case, we create custom pages to the many of those companies here, yep. So if you go to the Dallas factory, we have that custom page showing the assets that you can go to the area. If you go to the lich moose, you just go to their website. Go to Bosch. I just train some data tables because, you know they have so let's stay with our data that I know, but that's at the site level, just to the simulation that you can have in many sites system. Then we go to that, to the line. We have the KPI for the selected I'm not sorry, and you're not flying. So those are the many lines we have in that area. Then you can drill down to one specific operation for the line. So we put together dashboards that you can have traditional men. Control of states. You can have some historic information, some traditional operational data for that machine, and all live with your data. Then you go to a traditional OE page, and then if there is data that you don't understand, you just put a table with a data grids, and this cost, it can be applied to any kind of assets. Just for fun, we put out to some solar panels, yep. So the idea of this application is, as I'm navigating I have those different form, different types of assets that will react dynamically trying to have and if you want to see that application, iPad on iPhone. So I do a simulation for iPhone here. The scissor will automatically switch to a form factor that's more suitable to iPhone visualization, as you can see. So that's how your application looks on the iPhone, and this demo that are showing here is available as a web client through our main website. Also, yeah, I think
Speaker 4 21:08
Walker earlier was very excited to see what you guys have built, and was telling others to connect and just go to the IP address. Yes, you can view it from any phone, and you can buy so I think that's that's a really cool way to display the application, Harry, maybe to ask you on the same, similar train of thought, right when you're having these early conversations with some of the customers, and I think it's important still in our industry, I don't think everyone understands the value or maybe the insights of data. Maybe, what are some of the early proof of concept type of projects that you have seen? It could be maybe energy monitoring. Maybe it's OEE like, what other like conversations have you had? Yeah,
Speaker 3 21:44
the specific applications are all over the place with manufacturers. So we usually start with just base level data collection. Can you get some data in log? Some data to a historian? We work with Canary. We work natively with SQL, so you can log data to SQL to get some data collect, connect to your PLCs, get some data collecting, get some visuals around trending. We have real nice trending tools. That's a good place to start. Then you may move into maybe some downtime or some other energy monitoring, it's definitely an application that people are looking at. So and that one's pretty it's pretty straightforward. Again, it we've got great tools, easy to connect, easy to connect, easy to visualize. So get to that level, and then you can show it off a little bit. And we go from there, then we the platform, like Mark said, you could download it. It's we provide the development tools for free with two hours of run time, actual connection run time, the development will run as long as you need it to. And then we have a training site so you can go get some basic education on how to get going. And if you need help, from there, we have application people that can help out and tech support. So we got everything lined up to get guys going on their own, or if they need some help, you go to bar kids direct end user, but also through systems, Instagrammers. So that's a huge piece of what we're doing. We provide a great open tool per integrator to be able to take to his customers to solve problems. So end user, systems integrator. The other area is OEMs. So we grew up Mark has been working with OEMs for many years, so the platform lends itself to levels of customization that only OEMs, that OEMs have those specific requirements. We had a gentleman come to us less than an hour ago with an OEM type requirement, and then he was, okay, I need something to put on a machine that can handle the data, published, the communications, the graphics to publishing, and we're like, okay, we're this is a unique platform that you can put right on the machine and do all of your needs in one place.
Speaker 1 24:11
Very interesting. I want Vlad knows I love and the audience knows I love to ask forward looking questions. I love to get predictions. I'd love to know where you guys think the industry is going to go, and maybe share some of the new frameworks, version 10 features that you guys just really so I'd love to know what you guys release and where you think the industry is going to continue to go, either from a platform side or for a user need side, or any and all of those.
Speaker 2 24:38
I do one initial comment with lots of airs on that, about easy to use, okay? Because this is a good example. The data is fully dynamic. It's not an easy specification, and it took that only between one two weeks. So one important feature that must say. Panettua, previous question about the proof of concepts, it takes only from a few weeks. Worst cases are a few months, at very low cost, that proof of concepts, yep, the only reason may take a few months, it's not because of the subject configuration that you can do a few days. Yes, you decide what, yes, I want to do, and where is my data? After you have that that was flexible enough, and you cannot just a couple days to the wiring, that's one of thing that's important. Psychology. What's new professional time? So the first part of the question, what we put the effort, we put the effort to make more user friendly. We put the effort to be more flexible to deal with those dynamic scenarios that you are environment sharing. We put effort to be the best performance possible on web, graphics, okay, the market. We put effort to make the platform a future proof Yuka, meaning we put some infrastructure there for AI design that you collect the benefits over the next 510, years. What else on the feature sets that first sponsor look like? Yeah, I think
Speaker 3 26:17
a lot of them right there, the technology called refresh, bring it to today, future proof. It that was a huge part of version tech. Okay, so because we had been in the market for a while, we really needed to get to the new.net aid, or what used to be called.net Core, and also another huge technology upgrade. It's underneath the hood, but WebAssembly, it's definitely a game changer moving forward. So I see those as a big part of version time from a tech standpoint. Yeah,
Speaker 2 26:51
make use of Docker deployments, the plan, the air, the rocks of infrastructure technology, the second part of
Speaker 1 26:59
it, the second part of the question is, you guys have a bunch of super fun features. I love the comment of future proof. Where do you imagine that the industry is going to go? What will the client and user demands be? Where do you how do you predict the software will change and the needs of the clients will the needs of the users will change in the next two or three years, with
Speaker 2 27:23
Walker reigns will love to explain, I hope I quote him properly, even the absence is that the previous stack, we have all those stacked applications, yes, exchange data, well, of another One. And he put the concept to have the enterprise name space, and it's done. And put that together, that same concept you need to for not exercise using the concept of finding space. Think about any automation projects. There are many automation solutions that they do with too many building blocks. Yes, I saw some nice of course, not say exactly who somebody that was putting make proposal, put a nice gateway that, oh, I have this product here that can publish to this guy, that can publish to my gateway, and then my gateway can publish to this other guy. There are three complex and four parts involving how to move data to snowflake. How we move data to snowflake? We pick up any solution from the edge, low cost on the counter, from the high end, and you plug snowflake directly. Okay, so I think that our mindset, and believe the interview, will start to recognize that okay, a platform should be, deserve that name should be your generic toolbox that can solve all those requirements for data flow, okay, to normalization, data integration. And you should not be adding lots of components to few things that not bringing real additional value. Process, okay, you should be adding analytics, okay, you should be adding curse reports. You shouldn't be adding something specific that process, okay, but not having to worry interesting tools. Let
Speaker 1 29:23
me dig a little bit deeper there mark, because I Vlad and I have long talked how there are many companies in the space, and sometimes it's difficult to go pick the entire tech stack and as people, as a person who has gone through and architected and deployed a tech stack, and more frustrating as a person who sometimes comes in on the back end to see some of those choices. It does feel like we add complexity, right? There are many ways to go about doing a project, and it feels like some of them are maybe complex, just to be complex. Now, there are certainly projects who whatever reporting to there is an industry. Specific reporting tool, and that reporting tool makes a lot of sense than building a lot of reports. Or you're looking for one particular you're looking for one particular tool, or you're going to the cloud. Do you imagine that we'll see fewer companies, let's say, five years from now? Do you imagine that there'll be fewer competitors, fewer companies here, and there will be winners of, hey, we have gotten to the point of lots of people are using us for the entirety of our platform. Or do you imagine that the market will just be more fractured, and there might be 10 different ways and 10 different tech stacks that different customers pick and choose, higardvi
Speaker 2 30:38
specific skulls, yes, the number of companies will keep growing a lot, because you've talked about machine learning, reporting, madness. Those terms are completely generic, but how to implement in one specific sub sector is very specific. So you see a huge proliferation of companies working on this specific bringing the real value on the solutions we go for, the more infrastructure products you see that Mark shrinking. In fact, right now everyone knows that walkers does not measure his words. He talks what he thinks. Make that very clear he does not. Should be good or bad if anyone just pick out his mind. And I'm very happy to say that his knot, you say, okay, as platform, I seen at least now three players, that's frameworks and children. So even now, not top the long term, yes, way, at least for his point of view, where you need a wiring tool, he's coaching only three companies that are capable to do that infrastructure integration. And that's very important to understand our company. Times they have, oh, what you have all specific analytics. We have integration to any kind of company you can manage to provide the data easy way and present back. But our platform has really focused to be the infrastructure platform, whatever you should do to get data, move data, you recognize data, do calculation some data, do any kind of integration and any kind of advanced UI you need to have out of the box, what you can add? Harry, that's a lot
Speaker 3 32:36
for sure. I think the requirements are really all about data. So I think to your question about where do all the vendors and the fragmentation, I think always, there's gonna be some big players that emerge as kind of leaders for the big enterprise stuff. But I think there's, you know, where we're positioned. I think there's a place for us, and there it's a big world, you know what I mean. So I look at uns great, great concept unifying data that's happy that we met Walker and the kind of mark and Walker connected, and we built a lot of things around, around those concepts. So that was big stuff, you know, but I think there's a big world, even outside of prove it. And I think our platform is, the key is open, flexible, and then connected collect, it's the stuff we've been doing for years. Can I to collect, transform and then stay open, be able to fit in anybody in most landscapes,
Speaker 4 33:50
perfect. Let me ask you maybe a one last question. I
Unknown Speaker 33:54
was gonna say we've got about one last question worth that's
Speaker 4 33:56
fine. Last question before I let Dave wrap up, and I'm curious, as you're having these conversations again with customers, my feel is that there's still a struggle on this. You could label it ot it maybe integration or understanding more of these data insights. I'm curious to hear your thoughts on maybe, what are some of their challenges? I would say more on the personnel side right. Are they maybe lacking the people to be able to integrate such systems? Are they lacking maybe more data scientists? Are they lacking people who can understand both the plan for data, which I think is very critical, and then be able to translate it all the way up to it insights? What kind of conversations on that side are you having with with customers? We definitely
Speaker 2 34:39
have a people brown because they're completely structure inside the company that's very challenging, and they don't really have the property stand over the side. They don't talk much. But it's also it's a big brand that we try to solve. Our platform, men Tuesday life to say, and we allow OG it. Integration. That's a cop agency of the place. But that's not the reality in most platforms. So when you try to have that system, you go right away two problems. They have some legacy solutions that they're not able to play well, we used to try to cross domains and the people, they're not even the same. So everything's wrong, the people and the tools we are solving. The tools part, the people, Terry is a shoe
Speaker 3 35:31
is definitely a jealous and the amount of legacy technology that is out there, and I think with the newer push to IIoT or IoT. The ideal thing is for the IT guys to skip and just, hey, we can just connect right to these machines, and this will be easy to do, and then they start to run into the reality that you run into in an industrial, real time environment, yeah. So the good news is that I think there's a lot of great investment in the tools there, with platforms like ours that can help these guys bridge some of these gaps. One example
Speaker 2 36:11
is why you can run inside Cisco halters. Cisco halter the IQ can deploy, yeah, if an IPC, it must go through another batch Absolutely.
Speaker 4 36:23
And if I'd like really brief side note, last year, I learned that in it, a web page that is real time updates every 30 so that is nowhere near enough time for industrial and
Speaker 1 36:35
we also don't have enough time to get into the real time argument slide, because that is definitely more than 30 seconds of what is and what is not real time. Perry Mark, this has been absolutely fantastic. I want you guys to look at the camera. We're asking everyone one last question. If there's an end user who maybe didn't see you, who's here, if there's a listener, what do you want them to remember about tats off? What do you want them to remember about frameworks here at Pruvit?
Speaker 3 37:03
I'll go back to some of the things I was saying before, but think the open, flexible, designed with real time data in mind. Absolutely. The part I don't like to go to as much is very affordable,
Speaker 2 37:19
and I as its rates if the focus is on the solution, once rates not what this tool can do. Okay, lightwork pumps that are unwilling to compromise their specifications. So instead of learn the two and say, oh, what I can do that to know they want to achieve some results around real time data, and and we will deliver. So he put the focus in that solution, and they look in something that's flexible enough and with the performance that can deliver his specifications, so he does not have to think about the tool. Perfect.
Speaker 1 37:57
Mark, Harry, thank you guys so much for being here. I look forward to having you guys on the show again at some point soon, and seeing you're on a bunch of additional conferences this year. If you guys want to know more about tad soft, if you want to learn more about frameworks that we will now pronounce correctly, at least for today, go. Vlad does an amazing job. He's going to drop all of that information in the notes. We'll have it out on the videos, we'll also make sure that you guys can connect with both pets off as well as Mark and Harry, wherever and however you guys want to, want to go do that again, everyone. Thank you for watching manufacturing hubs continued coverage here at the Pruvit conference. If you guys want to watch more again, continue to check out these videos and watch Vlad and I normal Wednesday, four o'clock East Coast time every week as we continue to have these awesome conversations. Thank you guys.
