 Hey everyone, welcome back to theCUBE, the leader in live tech coverage covering supercomputing 23 from the Mile High City, Denver, Colorado, Lisa Martin with John Furrier. John, we're going to be talking about the data management revolution next and it's not just about storage. It's a whole nother ball game. AI has changed the game. This area and this interview is going to highlight what the impact of AI is going to bring to existing industries. It's going to be a revolution. Radical change is coming. We're going to hear a lot about data management here. Big impact. Please welcome Eric Basis, CEO of CalRay to the program. Eric, it's great to have you. Thanks for joining us. Thank you for inviting me. Talk about CalRay intelligent data processing from Cloud Edge. How is this different and unique from traditional processing? Data management, excuse me. Yes, so CalRay is one of the pioneers in the data management space, especially for AI and what we call data intensive workloads. We are a 15 years old company and actually originally we are a semiconductor company. We designed a new type of processor called a DPU and the reason for that was to process data more efficiently. And then little by little we realized that having just a DPU was not high enough for our customer to really leverage the power of DPU. So we built what we call the data management platform. And really the goal of the data management platform is to manage all the step of the workflow and make sure that you run that as efficiently as possible. So today, today we have about 100 customers in this space and a strong partnership with Dell. So if I get this right, you guys started as a semiconductor company, built a DPU data processing unit, CPU, DPU, all kinds of GPUs out there now for training. And then created a software platform that runs the chips. So actually our data management platform is composed of seven modules. Okay, the first module is actually what we call the engineer hub. So this is a data orchestrator, okay? The one who is locating data where you need the data to be very, very efficient. It sits on top of very, very efficient parallel file system because you need to access object of file very, very fast. And then below is what we call the NG box, which is a box with NVMe. And so our DPUs allows a very, very fast and cheap implementation of NVMe storage, actually. And that was news that was launched a couple hours ago. Correct, yes. The NG box, which is a common product with Dell, including integrating our DPU cards. And it's a pretty unique, fast, very, very, I mean, actually cost efficient implementation NVMe. We really appreciate you coming on and we appreciate Dell. We had, we were at the Dell HPC community event yesterday. And what you're seeing is this next generation of compute processors, combination of hardware, software, and semiconductors. Kind of like how PCs were basically, but now it's for platforms. And we're going into a platform engineering world where the game is still the same, but the parts are changing, getting smaller, faster, cheaper. What is the specific things that are happening that you guys are seeing that you're addressing and that Dell technology is addressing? That's different. Is it the complex workloads? Is it the more in data intensity? Is it both? Is it the old ways inadequate? What is the problem? What are you guys solving? The problem, I think, is very simple, actually. It's a problem of workflow. Now you need to think about workflow. What is your workflow? So there is a beginning, you need to capture the data, and then you need to process the data to go through different stages until you've got the result. And it's not so linear, actually. It's much more complex. So it's not so much like we used to be compute and storage and that's it. Now it's about how do you manage a workflow? And to manage a workflow that's complicated, you need to actually orchestrate the data movement. You need to make sure that the data are at the right time, at the right point. And you need also to have a fast access to the data, and for a wider and wider amount of data, while keeping being easy to use. So this is really a completely different scheme, a type of set of challenges, and that's the purpose of the data management platform, to manage all these challenges. Okay, go ahead. Thanks, John. Can you double click on the uniqueness of differentiators? You mentioned them, you kind of alluded to them with the news that came out today, but I'd love to understand more about what makes this offering so unique and how it's helping customers address the current challenges and the changes in the landscape. Of course. So I would tend to say that if you look at the main feature of the data management, the first one is the orchestration. Orchestration meaning to basically move the data where you need the data to be used, so that you don't have to wait for it, okay? And so that you don't have to duplicate all the data in your whole system. So first thing is really we've got a unique orchestrator who will recognize the industry, which will bring the performance you need from the beginning to the end. That's the first thing. The second key thing is about anyway, you need to have performance, so you need to access data very fast, okay? And here we've got a very unique way to access fast data using our Parallel 5 system, and which is also combined with our data management platform with our orchestrator. It's also a great way to have a better TCO. Let me explain why, because that's one of the beauty of data management. Not only it manages the workflow, but it also helps you to maximize the usage of a different type of storage you have and to minimize basically the fast capacity you have in your system, because the data management will automatically only use the capacity you need. Talk about the benefits with Dell, because Dell's been promoting this AI intense workload. You get the data intense workload. They're kind of the same. Data is AI, data feeds AI. So what is that use case with Dell? What specific data intense workflows and what AI intense workflows are you seeing that you guys are addressing? So back to what I said, when you need to run efficiently a workflow, you need to have different type of profile to support. It could be object, it could be S3, it could be file. And you need to have also different efficiency of storage. So of course, the beauty of working with Dell, we've got a great portfolio of storage solutions. And what we bring to this portfolio, we add the data management part and also the very fast tier zero that you need to basically optimize and run more efficiently the overall flow. So it's very complimentary with the existing Dell solutions. Talk a little bit more about the complementary nature. I was reading that the solutions are comprised of three components that are very complimentary. And how does that complementary nature help simplify things for customers in dynamic environments and in any industry? Yeah, so simplification brought by the solution Dell and Caller is actually very visible for our customer. Our customer today, they need for example to move the data from one location to another. It's a nightmare, okay? Thanks to our data orchestrator, it's very easy. You may have your data on premise, on the cloud, different cloud solutions, a nightmare. Very again, the data mover will help. We help the enterprise, whoever you are, to make the best and to facilitate the way you manipulate data, okay? And very again, using data technology and using data technology, you've got a complete solution to run this workflow which is very complex. You know Lisa, when we were last on theCUBE together talking about how the old way brings in the new way, replaces the old way, old way, new way. We're in a shift with AI and HPC here. You can see it clearly here, that this next level's here. And it's still unknown what's going to look like it's innovation happening everywhere. What is the old way of data management? And what's the new way? Because it's clear that always happens in a way like this. Every single inflection point I've seen since the PC that I've lived through, there's an old way and new way. And it's clear. What's your view on old way, new way? So my view of old way is pretty simple. I mean, you have a set of data, okay? You may have on-premise implementation, okay? With, for example, your GPU, if you want to run some AI training or inference, and then you run this set of data on your GPU, and you get the result, okay? You've got a little bit less old way is to do exactly the same, but only 100% on the cloud, okay? But a new way, but I would tend to say it's still old, right, just that you move the problem. I think at the end of the day it can be very expensive. You move the problem. What is the new way? The new way is that actually you have a certain number of different location, a site, a new company who needs to use this data, okay? So one site will get the data, locate data somewhere, okay? Another site may use it. And then you may have to send this data to the cloud to run the GPU for the training. And then you get back this data and you use this data for inference, for example, okay? So you see it's much more complex, right? And that's a way you have no choice. You have no choice. You need to go through all these stages and you need to leverage the different location. You need to leverage cloud. You need to leverage on-premise. You need to manage your different sites. You have no choice. And that's why you want the chip and the system on the same box because it could be at an edge location. Yeah, exactly, exactly. So back to, we are entering a hybrid mode, a hybrid world, okay? So the reason why we also provide chip and outdoor acceleration is that through this stage, you need also most of the time to have your own premise very efficient. And now for on-premise, that's what we announced. We propose these NG boxes so that our customer can have very, very efficient boxes and they can really speed up the whole process on their premise. What's the secret sauce that you guys have? Every company's got a secret sauce. What's your secret sauce? So I think the secret sauce is to really understand the system. The DNA of a company is semiconductor and is the media entertainment space. So we've been working for the last 10 years in the broadcasting market and it's about workflow. So we know very well how you manage workflow. We've been 10 years of experience so our secret sauce is really to manage workflow in very efficient way. Do you have a favorite customer example, Eric, that really shines the spotlight on the value proposition that Calray is delivering as well as with Dell and really what's in it for me as a customer? Yes, so we've got, of course, many different examples but I'm thinking one of them which is very interesting because it's very different from what you may usually see. The company is Geo-Technology and actually what they design is a way to analyze train, so on the edge. So they capture data, they analyze, they do proactive maintenance. So they collect all the data. They run AI type of algorithm just to detect if there is a problem with the train and the train can actually go through portals and automatically be scanned so that you can provide any monitoring data and check if there is any issue. So we did that with Dell, with NVIDIA and I think this is a very, very good example of a very innovative solution leveraging complex data management with our partner Dell and NVIDIA. So that trifecta of Calray, Dell and NVIDIA, was that catalyzed by customers to demand? Talk a little bit more about that and how you're really listening to your customers to develop the solutions that they need to solve problems. Very good question. I think now most of the customers when they come to us they come with a workflow. Okay, they don't, I mean the whole school will be okay and need this amount of capacity with this amount of performance but you know I would tend to say the new school would be okay, I've got this complex workflow. How can you help us to manage this workflow, right? Of course, course is important, but it's about workflow and in that case, I mean they tried to implement this workflow. They didn't manage, so they came to us and we've seen that with our workflow management with partner like Dell, with NVIDIA we were able to provide a complete solution. Nice. That's awesome and I think you guys are well positioned for this wave Dell has been revitalized with AI looking at their portfolio. They're going to have GPU emulation announcement I guess that we have started liquid ed announcement. They're kind of set up for this AI generation. By the way, it's all on prem too and it's not repatriation, it's net new. Correct. You can see more stuff being built. Exactly, awesome. What are some of the things that we can expect next from Calray, from Dell, from the partnerships that you guys have built? What should we be looking for? So really, I mean it's about sharing this vision on the market and making sure that we can deploy this vision more and more. We passionate about this new vision about workflow because you're much closer to the application of your customer. That's what excites me, right? It's not about, once again, it's much more than just capacity. It's about understanding your customer, understanding the problem of your customer and bringing a solution, which I think is very exciting. Absolutely, that's what it's about is those customer outcomes, the business impact. Last question for you, Eric. What are some of the things, I know Dell is in your booth, what are some of the things that the audience who's here can stop by at Calray booth and learn and see and feel and touch? So don't hesitate to come by our booth, 5-9-8. We can show you our new product. We can show you, we have great demo so they can show the complete roadmap and the workflow. So please come and visit us. Excellent, Eric, thank you so much for your time and shedding light on what Calray is doing, how you're really helping revolutionize data management for customers and helping them solve problems and allow them to use emerging technologies. We appreciate your time. Thank you. For our guests and for John Ferrier, I'm Lisa Martin. You're watching theCUBE live from SC23 in Denver, Colorado. Stick around, we'll be back with our next guest just a minute.