Transcript
Hi, Mike Matchett with Small World Big Data. We're going to talk today about a confluence of several of my favorite subjects. We've got object storage on the docket. We've got AI, we've got cryptography, we've got, uh, data sovereignty, data protection, data governance. It's all coming together. Uh, we've got Stefaan here with Akave. Welcome, Stefaan. Welcome. Thank you, Mike, for inviting me. All right. So you've put together a company here that has brought a lot of topics together for us. It's it's really I see it as a conversion. You don't talk about conversions, but I can talk about conversions. It's it's really taking some of the, the best things that we've, uh, seen in the past couple decades, really, and bringing, pulling them together. How did you get started, uh, with, uh, object storage first of all, so that's really the heart of what we're talking about. And then pull in a lot of these other things, these, these hot topics, you just didn't get railed down into doing an array. How did that happen? Yeah, I mean, look, I mean, I've worked in infrastructure for 20 years. I've done a startup in object storage before. It's one of the first startups in object storage back in 2010, right after Amazon came out with S3. We sort of like took that same approach for enterprises. I've also been in cryptography type of solutions before. I was on a startup in deduplication where we used cryptography to de-dupe data. So always been living in the data world, storage world. Some find it boring. I find it exciting. Um, and so we kind of started with coffee as like a new architecture that kind of breaks down that normal monolithic approach of a storage architecture where you have, you know, a software stack that manages storage. We sort of like broke it up into different components. And we took some of the unique primitives that we've learned from the past. Um, and so we kind of brought that forward in a, in a way that users have some flexibility and in offering or consuming SaaS, uh, or a hybrid model where they can run some of the software stack locally. Um, and the value prop that we bring to our customers is specifically the ability to move data more freely across neo clouds or existing clouds and their on premises compute, because some of the trends we have seen are one, uh, this trend of like, you know, this AI trend, which is very hot, right? But basically a lot more people want to run more queries, run experiments, lower their compute costs, lower their bandwidth cost, meaning the cost of transferring data because everyone's hit with the same challenges. When you move data of the cloud, you have to pay a lot of money, right? And so we kind of took that out of the equation. We don't charge for egress. It's a fixed price. Yeah. And you can move your data more freely to those new compute stacks. All right. So let's, let's look at let's look at the world here just for a second. We talk about object storage. We're talking about lots of data. Right. So this isn't just a simple like little cache memcache or anything like that. This is this is large stores of data. So things that would be stored in data lakes, data warehouses, video storage, streaming services, lots of stuff. Uh, and, but, and you mentioned cost, which is definitely something we want to come back around and talk about. Um, and obviously with object storage and some story about cryptography, we're going to hear about immutability because that's a really important thing. But I wanted to just step back and talk about this, this trend that we're hearing, um, that you actually told me about before recording about data sovereignty. What is tell me about data sovereignty? And is that something that's really a big thing for us globally? Or where does that come into play? Yeah. Data sovereignty is like the number one topic these days in the boardrooms. I would say when it specifically when you're outside of the US and it comes down to being more in control over your data. And so in the past, specifically in Europe, you were required to store the data locally within a European location. Right? So that was easy. But today things have changed to be really GDPR compliant, you have to also govern those controls, meaning the keys. And whoever runs the service needs to be run by a European entity. So, um, this is where we're seeing a shift where a lot of European entities are now looking at moving out of American clouds because of this conflict between the GDPR requirements and what we call the Cloud Act, which is a US enforced, government enforced act that allows us, the US government to access data sets stored with an American owned service provider, even if it's stored in Europe. And so with Covid, we're software first, right? So our goal as a startup is to enable others to run services and enable infrastructure for customers in locations where we don't even have presence. And so the idea there is that it's sort of a hybrid model where we are building the software and we are enabling others to run that software and offer local services. All right. We also run a service ourselves in the US, by the way. But yeah, we can talk about that. So yes, you've got you've got an object store service that people can can use here. Uh, and, uh, let's just one level down deep on that. So when I'm talking about data sovereignty these days, there's GDPR sort of regulations in Europe, but there's also this idea that if I get you right, the data needs to be where I am. So it can't be available to other foreign countries. Uh, the governance should be done in the same area, right? I shouldn't allow governance from other areas of the world, because that doesn't make any sense. And I should have the keys where I am as well. And that so. So if I take those three boxes and I go through a lot of other cloud storage schemes, they don't they can't check all three of those off very well. Yeah. It's hard or you have to pay extra for it. Yet you have Amazon, Google and Microsoft. They're building sovereign clouds in Europe right now and they charge more for it. Oh, and they work with local integrators. But I mean these are massive integrators obviously. Where where we see interest from customers is customers that want more localized. When I say localized, you know, close to where they operate their services and there's not always a hyperscaler cloud available. So we're kind of going back to the days where the MSPs or the, you know, value added resellers and integrators, those local, more localized integrators have more of the customer base. And so we are seeing more geo patriation from the cloud For the reasons that I mentioned. One is sort of like the need to reduce costs, right? Because you want to run more compute. So you look at alternatives. Then you realize, oh, my data is kind of locked in to the sovereignty, right? Oh, I need to be more compliant. What am I going to do about that? And three is sort of like this new wave of like AI readiness, like AI agenda, AI workloads. What can we do differently now that we have a new architecture that would support the new, uh, consumers of resources and that's going to be those agents. It's still nascent, but it's going to happen over time. And so having infrastructure that scales and is fast in, in responding to these agents requests, that's something that we also focus on. All right. So we're designed for sovereignty. Uh, just just quickly you can help, I think indicated you can help people trim costs because you're going to bring things locally. You're going to enable local storage, sort of this hybrid approach is really about cost control and, performance. Bring the data close to where it is. And then resiliency. Uh, because of the built in. Uh, erasure coding. Erasure, coding, cryptography. And so, yeah, you're right, you're right. I forgot the most important thing to. Talk about is there's cryptography built in in the sense that for every data. So it's an S3 compatible interface. You can use your existing applications, but under the hood, we are calculating all the cryptographic hashes that are based on the content itself. And those get stored on an immutable ledger so that you can guarantee data provenance, um, of what data was stored when, when it was deleted. So you have a full audit trail that is immutable. And that seems to be even more important, not just in the backup and data protection world, but also in these newer AI, um, genetic workloads. And we also see that in the sovereign, like heavily regulated environments. Actually, we didn't talk about this much earlier, but, uh, we've seen a research in the research, uh, community. We've also seen that NIST 801 71, I believe is like a regulation that is being more for more, more aggressively enforced. And so that requires actually more auditability on where the data is stored even in the US, right? And who has access to it because of security concerns. And so we're going to see that more and more. Um, and this is where we can help with. All right, so, uh, we've got to kind of cut off here, uh, which is just the tip of the tip of the iceberg. And that's another topic we could talk about is iceberg and parquet files. Uh, but, uh, tell us, uh, you've got, you've got, uh, some apps out there that are using this. You've got protection vendors, partners already that people know, uh, you've got, uh, ways to work with, you know, I just teased that out with snowflake and some other folks. Um, if someone wants to find out more information about how you're bringing all these technologies together, how you're being able to make this a more cost effective service, how you're enabling geo patriation, repatriation of objects, storage, large files, and making it cheaper. Where would you send them to? What would you have them look at? Yeah. So we have to answer. In short, we have we're qualified with snowflake. So you can run your compute still in the cloud with snowflake, but store your data with a cave and, you know, move it to any other compute where you want to, um, to, uh, Veeam, Acronis, right? The backup layers. Uh, we also have some streaming applications that are integrated, uh, rclone sort of like the standard tooling for DevOps. And, uh, if you want to test it out, our SaaS offering right now is easily accessible at Akave.com. Okay. Where you can register, you know, you get some free storage for two, two weeks. Um, and then we'll just, uh, you can also reach out to us. There's a form on our website, but you can start playing around with it. It's easy to connect your applications. And if you have more questions around like the X4 oh two, which is specifically focused on agents, which is a payment wall where our MCP server or anything else that you want to talk about, let us know because we're here. We also help a ton of developers that are using our SDK. Lots of other stuff to talk about, but I'll leave it at that. Just had to drop that at the end. And oh, by the way, we have a paywall built in to help with Agentic AI workloads. You just had to drop that in there and we got to go. Uh, come back and come back and have a let's talk about that for half an hour. All right, let's do that. Uh, and, and get into the AI part about this. Uh, thank you for being here. Uh, I'm just really excited to see object storage reinvigorated again and coming around with something quite interesting to, I think most people out there. So thanks. Thank you, thank you. Thanks, Mike. All right. Take care.
