 Well, welcome back to SuperCloud 3. I'm John Furrier, host of theCUBE. We're here for a keynote conversation with Russ Davis, chief operating officer and chief product officer of vicinity. We're going to talk about more and more data being seen in more places needed for more applications, opportunities for this next gen cloud, the super cloud movement that's happening. And Russ, thanks for coming on theCUBE for this super cloud conversation. Thanks for having me, John. You guys are doing some really interesting things around data and this next gen kind of workflow we're seeing where applications are using data. Super compelling, it kind of points to the super cloud trajectory and what people are thinking about. But before we get into it, talk about your company. What do you guys do? You guys are in an area that's super growth right now in data. Give a quick plug for what you guys are doing and introduce yourself. Sure, so Russ Davis, as you said, I'm the COO and CPO for vicinity. vicinity was actually created in 2018. It is the reincarnation of a prior company called Bay Microsystems that was primarily focused on the federal government. That's where the original funding to develop our technology came from. This place is like the DOD in the intelligence community. vicinity basically was to take that commercial and start to apply the technology that fundamentally does one of two things. One is we move data across wide area networks faster than any other technology. Because of how we do that, we're able to actually enable compute to work with data that is remote from it. And I know that seems like a misnomer of some kind, but for many applications that actually works. And if for whatever reason, the network isn't big enough or there's just too much data, we go back to the first side of our coin, right? Which is we just then move it to where the compute is. So two tracks there, two sides of the coin, so move data to where the app needs to go really fast and cost effectively what people care about. We'll get that in a second. Or move the compute to the data where it lives. That's really the two approaches, right? Fundamentally, yes. We're basically enabling our customers to utilize their wide area networks almost as part of their storage fabric, right? So we're extending that storage and the data that's sitting there to the compute wherever it sits. So you guys are perfect company to point to when you talk about SuperCloud, when we say SuperCloud as this new multi-cloud kind of environment exists, but also it's not fully optimized. And we're seeing kind of this distributed computing mindset come in with new stuff. We got a tailwind with AI and machine learning. You have new applications, the development and open source is booming. So you're going to see an accelerated functionality in this new SuperCloud. What do you see for SuperCloud? What's your view? How do you define SuperCloud? Why do you like it? So for us, it's great, right? I mean, look, the clouds really became a big thing because of this whole idea of scale out compute on demand. That's kind of what got them to where they were. And then it was the idea of cheap storage, whether that's true or not. It's a different story. But then we see hybrid cloud, multi-cloud. We hear all these different names being applied to what? It's enterprises or companies in general just looking for how do I optimize my IT infrastructure? And for a company like vicinity, it's literally almost like mana from heaven to quote my CEO. Because there's so much data being generated in so many places. How do we process it? Where are we going to process it? And people are making those decisions based on the relationships with different cloud providers. But then maybe because of an acquisition, I've got stuff at Azure, I've got stuff at AWS. Oh, now I've got stuff in GCP. How do I make all of that work together? Well, what's the value of the work? It's insight from data for one part of a business process or another. We're able to connect those. We don't care which cloud it's in, whether that's public, private, a colo, on-prem, doesn't matter to us. Our whole notion as a technology is to connect the computer, the users, to data wherever it sits. So the question that jumps in my head is, one, how hard it is to do that? It's a hard problem. And two, costs come up a lot. So I hear people, that's the nirvana. I'd love to just have my data accessible, highly available, and with high availability, kind of multiple perspectives. But it's a hard problem to solve. It is. And it can be costly. How do you respond to that? Well, I mean, the big thing is that you're going to have to store the data somewhere. And so, basically, on what is the most effective and from a cost, but also from a compute, right? How long does it take to process data? I mean, traditionally it's been, my data's got to be with the compute. So wherever that is, I have to pay for that storage, right? I also have to then pay to move it to wherever that location is. I think what we add to that equation is put it anywhere, and you won't care. You don't want to have to egress data from a cloud provider, right? Paying by the drip is not a way to do that. So if you put it up there, you better be able to keep it there and only have to take out a little bit at a time as needed, right? But what we see is a trend where more and more are actually deploying storage on-prem but using burst compute to the cloud, right? And that could be any of the number of clouds, right? So this is the product I want to get into. So talk about the vicinity product. How do people engage? How do they deploy it? How are they managing it today? How do you guys solve that problem? Because again, it is nirvana. I can see that being in high demand. I want to have a nice, clean environment for my data. I want it to grow. I want it to be useful. So what's the product? How does it deploy and how do people consume? Sure, so we have a couple of products, right? But fundamentally it is software that's available in, say the AWS or Azure Marketplace. We run in GCP. We're not in the marketplace, per se. And then that allows for us to do software-only scale out types of solutions. And then we've got actual hardware assist technologies to scale up. So when people think about moving data fast over networks, they think, you know, when optimizers, extreme file transfer technologies, none of them go up to 100 gig. We do, right? And we do that using off-the-shelf PCI cards that are manufactured commercial product because they have FPGAs on them that we can leverage. And one of the benefits to that is that we're also now seeing a movement in the cloud providers, especially AWS and Azure, with exposing FPGA-based instances so that you can actually have specialized compute. Well, we leverage that to do specialized acceleration for data going into or out of those clouds. Now the big thing right now that everyone's working on right now that's taking all the applications by storm is the app developer market, AI and machine learning. Generative AI in particular is transforming all businesses. How do you guys fit into that trend? How is that going for you guys? What's in it for the customer? How do you pray that? Because, again, data is critical for this workload, if we're training and doing inference. Sure. So, I mean, our origin was actually in high performance computing, right? We originally did fabric extenders for InfiniBand networks to move, again, data over very long distances in big pipes. So, what we see in the trend today is that a lot of those applications require that that data be moved to a large compute farm, right? And to do that, it's a lot of data. And in a lot of cases, it's in different locations. So imagine if you could say, have data generated in five different spots that you wanna run some AI workload against, but you could do data fusion in real time because we can connect that compute to each of those different remote locations. Real quick, display data fusion for the folks watching. Oh, data fusion in the sense that for my algorithm to be effective, right? Because I'm doing some kind of, I don't know whether it's, you know, graphical neural networking or whatever I'm working on. I actually have data that is not in one place right now. But I need to actually analyze all of that and bring it together, fuse that data to be able to get an effective insight from that data. So as I work my model, my model has to account for different. Well, I wanted to just call that out because I think that is one of the trends that we're seeing the most of now. Using data at the right place at the right time is as part of the developer and the application. Correct. To do something that it was, that either it would be magical in an AI sense or something that's needed for the application, which by the way, it's hard to do. So that's a, I'm going to see more of that. Absolutely. And I mean, I just recently was working with someone who was debating using something on their home computer versus do I use a co-lab on Google, on GCP, right? Who gives you a free GPU instance to work with as a data developer, right? And so I think we're just going to see that grow more and more where we're going to have, you know, these GPU farms that are not going to have the data local though. What are they going to do? Yeah, I think, I mean, I think chatGPT was a great educator for everyone in the world to see what AI could do as a use case and the frequency of that, of that regenerate this result. That's going to happen under the covers in the application. That's why data fusion is super, super important. And so first we're kind of in this evolution of let's get multi-cloud set up, let's get hybrid cloud set up. Now you got, that's like more platforms of service and with underlying infrastructure abstracted away. Now you have this next layer, the super cloud layer that's going to do smart things. That's where the action is right now. How do you respond to that? And what is vicinity doing there? The big thing for us is ensuring portability across all the different cloud architectures if you will. Because A is not like B is not like C kind of thing, right? And anyone who's worked with AWS or Azure or Google, they know. So how do we focus as we're doing the data developers, knowing that I want to be able to work across these different infrastructures. So for us, and now for my product side of things, it's really about making products available that can actually touch each and one of those. Now remember, and for us, it's about focusing on data accessibility, right? We're not trying to do data management. We're not trying to do some of the other more elaborate things that workflow management systems, MAMS, DAMS, things like that do, right? But it is all about making sure we can access whatever is that next big infrastructure, right? And that's where the management comes in. It's works together with what's available on a governance basis. And then you're making it accessible for stuff, infrastructure, software, apps to integrate with. Because we sit below the application layer with our technology, we don't care what the application is, right? It's a novel position to be in actually. So of course we love the AI machine learning line. We think that's going to be a tsunami of new innovation, new creativity. Certainly as data becomes accessible, there's going to be more new things emerging that we could never see before. So I think that this data openness is going to be a big thing. And thanks to you guys and others. But the question that comes up is, okay, you got me at accessibility. Okay, I love that. Now, security, the fear kicks in. How do we make it secure? What's the data security angle? Well, there's a couple of pieces to that, right? One is, as I said, we sit below the application layer, right? So we don't touch data. One of the interesting things I mentioned, how fast we move data. But we don't compress it or de-dupe it, right? Unlike other WAN optimizer types of technologies. Think about what I said about our origins, right? DoD and IC world. They would give us encrypted data. Can't touch it. So we had to figure out how to accelerate across the networks without ever munging on the data. And that's what we've done. However, what we do is, we were actually FIPS 140-2 certified in our transport across the wide area network. So we use standardized encryptions, right? We're literally protected there. And we've got some other deeper technologies I probably won't go too deep into. But we have the ability to do data obfuscations and data across different paths, encrypt them uniquely. But we don't do anything that affects the rights that have already been assigned to that data at the user level. So things like ACLs and things that we don't touch them, right? So whatever that is from the enterprise management side of things, if they're using AD or LDAP or something like that, we won't actually touch it. So you basically work with whatever existing security protocols are in place, but also have your own. What about data democratization and interoperability? Obviously we're seeing a democratization of the revolution of data as it gets refactored, if you will. Snowflake Databricks are two examples out there. Others, working with multiple data providers is something that is going to come up a lot. How do you see that evolving? And how do you answer that super cloud data interoperability question? Well, the big one that I've seen so far has been organizations wanting to leverage object data that's sitting in S3 buckets, almost like they do file data, right? In fact, it's surprising me to hear that, object storage was designed to be basically cheap and deep, right? I mean, that was the concept. And so there was so much invested in that, but now people are realizing, oh my gosh, I put all this data into an S3 bucket and I now need it. I need it like I do file. So one of our big things is vicinity fully supports being able to go from file to object, object to file. We have a cloud native technology that we actually just announced about, I don't know, six weeks ago with Teradata to support their new network object storage, right? So that you can run cloud compute against an object store sitting on prem, right? Without having to do that. So we're working towards how do we change formats, but fundamentally within the unstructured data world, right? Because when you start talking about the structured data world, the different database formats, there are other technologies out there that deal with a lot of that, right? But the problem in that arena is that they're all proprietary. Russ, it's great chatting with you on your product of accessibility. It's one of our core pillars of super cloud and that layer. Talk about vicinity's value proposition. Where are you guys succeeding? Can you give some examples of real life examples of how you're being deployed? What are the use cases? Because you've got the pioneers now probably using this and progressively pushing new architectures. It's pretty cool object and block, having all that interoperability in this abstraction. Where are you winning? Where is it working? Can you share an example and just anecdotal data? I can give you a couple of them real quick. Big thing in terms of our value proposition are those that need to derive value from their data in a timely manner, right? So think time to insight, time to decision, time to act. So the federal government is obviously a very big customer of ours. And then on the commercial side, I'll give you a couple of quick ones. One of the big ones we just did with a cruise line was they needed to update servers on a ship at sea. Satellites, low bandwidth, high latency. They literally took a three gig file, moved it from shore to ship over that satellite. It took 16 hours. With us, it took 27 minutes. This is a huge difference, right? We actually just- A ship's like a floating data center, basically. Absolutely is. It absolutely is. People probably have no clue just how much technology is on a cruise ship. That's a story I would work on with the team. We're gonna go hit that cruise ship and do a deep dive. I'm happy to come with you. Happy to come with you. Okay, so continue. And so actually just data movement. As an example, we did a webinar with AWS within the last two weeks here where they wanted to, they tested using our technology versus AWS, whatever they threw at it, to move various types of file formats, sizes, right? And so at the peak, doing very small files, like in the kilobyte range, we were 70 times faster than anything they were doing. In the reality, the real world, mostly people have mixed small files, large files, whatever. They're basically about 12 times faster, right? This is public. This was in a webinar. But the biggest thing when I look at the people that wanna do this, especially AIML kind of stuff remotely, we did an oil and gas concern between Brazil and Houston where they moved one terabyte of a subsea scan over a two gig link with another vendor, I won't name. And it took them just over 16 hours and 15 minutes roughly, right? And that's not bad. They used us over a one gig link and it took two hours and 15 minutes. But what they did after that was they ran the application from Brazil against the data that they had just moved to Houston over that same one gig link. And they ran the application and got their result in 18 minutes. Why move the data? Yeah, this is exactly the point. I think you bring optionality to the companies deploying large data and changing data analysis. Look at how hard it is to get GPUs these days or stand up, say physical infrastructure for AI, for instance. I won't say who told me this, but I was at an event recently where someone told me that they did a number in the bees on a pre-buy, prepaid to NVIDIA to get allocation because the allocation is that tight. Yeah, I think we reported on that story. But anyway, this is the point. The supply chain on hardware, Amazon's got some nice sustainability goals. The clouds are working at scale. So the cost to push it and get benefit, if done right, is a new architecture. It's not like it was around before. So the old model of data movement was, well, let's not move the data. It's too expensive and out of the cloud, you can just keep it there. That's what you guys are saying. We don't care where it's at. You can literally keep it on prem and your colo in the cloud. Doesn't actually matter to us. Bottom line for the folks watching, what's super cloud going to look like in a year or two, five years? Now that you have an accelerated AI push, which will force applications to the market, I think we're going to see both the physical layer get smarter, obviously, at the hardware level. But as the super apps are coming out, that have AI built-in, generative AI and other cool things, we're expecting an acceleration. What do you expect to see the super cloud architecture evolve to? Wow, so it's an interesting question. I expect to see it extend all the way to the far edge. I fully expect that what we think of as cloud today are large data centers, city in Ashburn or Oregon or wherever, or here in the valley. There is going to be such heavy compute pushed out to edges, simply because there's going to be data that needs processing in microseconds. But then there's also going to be the larger result of that that needs to be fused, combined, and assessed that's going to take hours, weeks, months. So I think what we're going to find is with the decreasing cost for bandwidth, you're going to start seeing those connections go further and further out. It's interesting, super computing mission, high performance computing, HPC earlier. That's been around for a while, very expensive by the way to do. I think super clouds going to be things that are going to, we're hard wants to do and expensive to do are going to get in more inexpensive, not cheaper, but more less cost. And that's going to enable new things to happen. For example, real time data. As real time becomes important, you're going to start to see that mixed with analytical previous data. So fusion's going to happen. So I think more of that's going to happen at scale, which was hard to do and expensive just 10 years ago. Very much so. I mean, if you think about what's even happening in say retail for example, computer vision has changed everything. Now I want to deploy at least enough compute to my store, especially a super store, where I can do analytics against live video in real time locally, but I still want to need to process all of that data for other purposes, right? For bigger marketing or security, pick your subject, right? Russ, great to have you on SuperCloud 3, keynote conversation. Thanks for coming on, I appreciate it. It's a pleasure. Thank you for having me, John. This is theCUBE, SuperCloud 3, keynote conversation with Russ Davis, Chief Operating Officer and Chief Product Officer of vicinity, changing the game, making data available. This is what developers will want. This is the future. This is what SuperCloud will emerge and enable. And you get all here in the CUBE coverage. We'll be back with more coverage of SuperCloud 3 event after the short break.