 Welcome to theCUBE's coverage of Dell Technologies 2023 from Las Vegas. And prior to the show, we took the opportunity to dig into a really important topic, railroad safety, which along with rail car maintenance has been in the news lately. And with me are Ali Beers, who's the director of product marketing for Edge Solutions at Dell Technologies. And Jeff Nitchai, who's the chief technology officer at Duo's Technologies, really interesting company. Folks, welcome to theCUBE. Thanks so much for coming on. Thank you. Thank you. Jeff, Duo's Technologies, tell us about the company, really fascinating work that you're doing. Sure, Dave. We develop and install comprehensive rail car inspection portals, which provide a 360 degree view of the train as it's traveling at track speed. It could be up to about 125 miles an hour. Now, from that, we create high resolution detailed images that are acquired of the top, the sides, and even the undercarriage of the rail car. Now, we use artificial intelligence on those images to automatically identify defects and anomalies on those rail cars. We present all of that to rail car inspectors, which basically makes their job easier. It makes inspection faster and more accurate. So in a nutshell, that's what Duo's does and that's what the rail car inspection portal does. It's pretty amazing for the audience. If you Google this topic, Google Duo's, there's a really cool video on Facebook and it shows the train going down the track, like you said, Jeff, 125 miles an hour. And it's almost like it's going through like a mini car wash that's really high with no scrubbers. And so all this happens in like an instant. How does it actually happen? What kind of technology is involved to make this work? Yeah, you know, you mentioned car wash. Actually, some of our customers refer to it as a car wash because a lot of folks think it looks like a car wash. And I guess in this case, it's kind of looks like a train wash. But the rail car inspection portal itself is assembled from the ground up. It utilizes a custom built truss system and a steel canopy over the track, which gives it that look, you know, you have a steel building over the track. Now the canopy provides a controlled environment for imaging and some protection from outside elements, like the weather. Now underneath the canopy is where the action really happens. It's where the magic happens. And that includes an array of line scan and area scan cameras, precision optics and lighting systems of lasers and track based linear speed sensors and AI tag readers and more. So there's a lot of technology packed under that steel canopy. Now all of this equipment is, it's powered and monitored from a trackside edge data center where all of the servers and the power equipment are located. So really if you want to know how it works, you know, as the train passes, those mini cameras and sensors and related devices, they scan the train and perform that high resolution imaging. That image, the images and data, they're recorded and stored and it's actually processed in real time at track speed. But, you know, the process itself and the processing itself goes far beyond simple image acquisition. Like I said, there's a lot of technology being utilized. We utilize line scan cameras for high resolution. So the cameras themselves have to be synchronized with the speed of the moving train. In addition to get a usable image of a rail car, we have to undergo a process called car cutting, which, you know, it gives you a usable image of each individual rail car independent of the entire train. So there's a lot of stuff, a lot of processing that goes on. And, you know, it's all done there under that car wash head. So that's kind of how it happens. This is unbelievable. So there's a huge change, if I understand it. So prior to this, you had inspectors so the train would have to stop, slow down. I mean, stop completely for, I don't know how long, but give us the before and the after. What's the kind of the ultimate impact that this is having? Yeah, honestly, you know, if you're not using remote and, you know, AI assisted rail car inspection, what your rail car inspectors are doing is they're walking the track. Usually they have two or three of them. They're walking alongside the train and they're crawling in and under, you know, under the rail car to see the underside. So there's a big safety aspect to it, as well as a speed aspect. You mentioned the train being stopped. Of course, if I'm gonna inspect the train physically, that train has to be stopped and that's what we call dwell time. And so what we're trying to do for the railroads, obviously is eliminate or limit dwell time as much as possible. So most of the anomalies can be detected remotely, just using visualization and thereby, you know, decrease dwell time, increase safety. And that's the target of the solution. So there's a big before and after. So, yeah. Very amazing. Hey, you know, Ali, I feel like we've been looking forward to and waiting for this day for a long time. And you guys talk a lot about delivering real-time insights at the edge. Other companies obviously do as well. This is pretty real-time, isn't it? Yeah. You know, what's your take on this? I'm a mess. I mean, unbelievable. This is just an amazing example of an edge use case. I mean, I think often when we talk about the edge, we're talking about use cases, we're using the word business outcomes, but what it really means when you're delivering real-time insights is that you can go from something that used to be manual into a more automated fashion. What it means is you might have top and bottom line revenue impact and you're gonna have increased accuracy and all these other time savings, well time, that Jeff was just talking about, but in the case of Duos, not only are they improving this rail car and employee safety and enhancing the accuracy of some of their measurements, but this can also help them avoid accidents and ultimately potentially save lives and really just doesn't get more important than that. So this makes sense, of course, because the industry, you got a challenging environment, it's critical infrastructure and it's harsh, but like what's inside? I mean, it got compute, you got storage, you got sensors, you got cameras, paint a picture for us, Jeff, if you would. You know, obviously Dell has been instrumental in partnering with us for the last three years. I'm sorry, for the last five years of the solution and most recently, we've kind of changed up the solution to have faster kinds of servers and storage mechanisms, but we have an array of servers and an array of networking equipment that we have in there for the past, like I said, for the past three years, we recently upgraded so that we can do 125 miles an hour. Now what that represents, just so you know, for two tracks, trains moving roughly 120 miles an hour, that's a little bit of over 80 gigabytes per second, which is an enormous amount of data to be able to store and process. So, you know, we certainly rely on the compute power and the storage mechanisms that, you know, Dell helped us bring to fruition and design the mechanisms that we need to be able to do that kind of speed. Yeah, and it's not just across this industry. I mean, I think this particular use case really highlights the paramount importance of safety and security, but for computer vision, it really does have an impact across multiple industries. If you think about a retail store or grocery store, when something falls off the shelf, somebody can be alerted and before a slip and fall accident can occur by somebody in the store, computer vision can be used to spot that. Same thing for manufacturing, somebody can be spotted when they're working too close to a robotic arm or a machine and it can help spot and identify those to make it a far better and safer experience. So, I think just technology at the edge done right can help deliver incredibly powerful insights to help you run your business better, but also to deliver these incredible outcomes and real innovation. So it's exciting to work with customers from across all industries to make these innovations possible. Yeah, I'll bet, I mean, I'm curious as to what's inside, what's a trackside data center? I don't think I've ever, well, I guess I, I guess some of the race car drivers, you know, they pump data into a trackside data center, but what does a railroad trackside data center look like? So the trackside data center for us, we call it the edge data center because again, all of the processing and storage is being done at the edge in real time because the results are required in real time. If I'm going to inspect a train and inspect rail cars for safety issues, I want them as quickly as possible. So preferably while that train is moving, right? So we do everything in real time or close to real time, the edge data center, which resides trackside is absolutely necessary to do all that processing on site. So inside the edge data center, which some of our customers refer to as a bungalow, but really it's a separate data center, it's a separate server room, right? And so what we have is racks and racks of servers, some dedicated to storage and some dedicated to monitoring and control and some just hosting our software that the customers use. And so that array of servers and storage technology is right there and obviously it's all environmentally controlled and secured the way it should be. So we don't have to worry about backhauling information to be used immediately back to a cloud data center because obviously the size of the images and the size of the data that we're backhauling would be way too large to get it in a timely fashion. Dali, I'm interested in sort of Dell's role and how you think about it because when you first go back to the beginning of Dell saying, hey, we should do something at the edge, you had a choice, you could think about, okay, we're going to apply horizontal technologies, but you know at the end of the day it's going to be industry specific and I think you chose to go, obviously they made the right call with horizontal technologies, but here's a perfect example of you're applying it within a specific industry. So I'm curious as to how you think about helping various industries and how applicable this is horizontally. Yeah, there are so many industries now that are requiring immediate access to the data and insights and that does often require processing the data right there at the edge. And when you begin to think about doing this at scale with many locations or distributed geographic footprint across multiple locations, this can get incredibly complex. And so while we've been working with customers at the edge for about 20 years and we have a very sizable business that we've been helping customers with for a long time, we're looking to continue to grow that and to make it easier for customers. And what we see from customers is really three common things that they're generally trying to reduce the things are still too complex. So reduce the complexity that they're seeing. So gain more speed, be able to do it faster and better and less manual. The second one is to do it in a more secure way. And then the third is really to be able to do it at scale. So if you think about why this is required and how our strategy came about, it's because the edge is the exact opposite of the controlled environment that you would find in a data center or a cloud. The edge is geographically distributed and it doesn't always have a qualified IT person sitting right there. So as you're standing that all up, you need to really design it with the end in mind. Otherwise you're going to grow use case by use case and stack by stack. And you're going to end up with a lot of snowflake solutions that are really difficult to manage and secure in math. So our goal for our customers is really to help them simplify their edge. And we do that with the best in class infrastructure that Jeff was referencing. We also do it though with our expertise in vertical solutions, much like you were just talking about. So we have practice experts that specialize in particular industries that come to the table with a depth and breadth of knowledge to help our customers through particular industry specific situation. And we actually build also validated solutions and designs so that our customers can have a faster time to value when they're trying to put solutions together. And our latest evolution of that is that we're continuing to help customers innovate and to do things in a less manual fashion with project frontier and the announcements that we're making after technologies world around that. So we're really excited with our Edge operation software platform to help customers do some of those things to help them simplify their edge in a much easier way. It's so cool to see this come to fruition because we've been talking about it for a while and to see it in real life. Allie, it's pretty remarkable that I think what you can do with a couple of cameras maybe more than a couple, but still you got some sensors and it can tell you what's happening on this multi-ton railroad car. Jeff was talking about bits, I caught thinking in my head, bits slicing to get to the different cars and really understand each individual car on the train. It's pretty mind boggling how much technology is sitting right here all at the edge as an independent entity, right? I mean, you're doing all the processing there, is that correct? Yeah, that's true for sure. And Jeff, you can speak to it as well, but I think what's really amazing about it is that you can begin to see those data and insights in aggregate no matter where that infrastructure is deployed. So of course we're talking about one specific instance at the track side, but of course you end up with a lot of those locations. So providing customers with access to their data in a real-time fashion really creates game-changing opportunities for them to unlock new innovation and see trends and make decisions that they never knew possible before. Do you persist the data? Do you send any of the data back for analysis? Well, how's that all work? So we do. I mentioned AI. We have artificial intelligence, machine learning models that we use to detect anomalies and any kind of defects on a rail car. So what happens is a lot of the results from that are verified in real-time, both by rail car inspectors, as well as by a team of data scientists and some folks that we have back at our office here in Jacksonville in our 24-7 data center. So they're looking at those results in real-time. And what we do is we have a continuous learning approach. Basically, we feed those results back through the data science portion of the artificial intelligence to make those models better over time. So that's one of the ways that we use, I think, the aggregated data from all of our rail car inspection portals. And the other way, obviously, again, I mentioned 24-7 support and monitoring for these critical systems. It's done. We reach out to every single one. We're monitoring every single one in 24-7 capacity. And so now as far as data, there are a certain portion of the data that we do aggregate to our data center here, but it's mostly for analytical purposes. So the images, that's the big buckets of data that have to remain on track side simply because we don't need to aggregate them here. They remain in those edge data centers. I hope they answered the question. I hear you. You don't want to push all those images through some thin pipe. You'd have to put them on some kind of mobile device and ship them somewhere. What's the point? So is this worldwide, nationwide? Where is it deployed today? And where is it headed? So we have 13 portals and they're deployed in North America. So both here in domestic US as well as Canada. And we also have a couple in Mexico as well. So it's here, we're looking to go international. We are talking to folks internationally. And so we know that it's gonna grow because there's been a lot of interest in it and folks are getting very, very serious about it. So we see the benefits. I mean, the benefits are real and they're undeniable. So I'll give you the last word. How do you see the future? I mean, obviously we have this use case that is very specific to rail, but it sounds like there's so much opportunity here for Dell and your customers to think about the future. And I think we're just at the beginning of really being able to unlock so much innovation. So this is being driven by the fact that compute is smaller and less expensive and that you can put sensors, small sensors now into so many different things. So I feel like we're just at the beginning of this chapter and our goal will be to help customers to do that in a scalable way over time where they can have the ability to manage and secure and lifecycle all of their compute that sits out there, whether it's track side in a retail store on a manufacturing floor or otherwise to help customers. We've been, like I said, we've been just waiting for this day for years. We've written about and talked about most of the modeling, most of the AI work today is done in the cloud for modeling, but AI inferencing at the edge is going to explode and we give a bunch of examples. This is not one that we predicted, but it's a good one. And I think there are many more to come. Guys, thanks so much, Allie and Jeff for coming to theCUBE, it was great to have you on. Thank you. Thank you so much. All right, keep it right there for more coverage from Dell Technologies World 2023 in Las Vegas, you're watching theCUBE.