In The Arena by TechArena - Ocient: Revolutionizing Data Warehousing for AI Workloads

Episode Date: October 17, 2024

Live from OCP Summit 2024, this Data Insights podcast explores how Ocient’s innovative platform is optimizing compute-intensive data workloads, delivering efficiency, cost savings, and sustainability.

Welcome to the Tech Arena, featuring authentic discussions between tech's leading innovators and our host, Alison Klein. Now, let's step into the arena. Welcome to the arena. My name is Alison Klein, and today we're coming to you from the OCP Summit in San Jose. And it's another Data Insights podcast, which means Janice Mirowski is here with me. Welcome to the program, Janice. Oh, thank you, Alison. It's great to be

back again. Janice, I don't think we've ever podcasted this much in a couple of days. It's incredible. I'm having a great time. Same. It's exciting. And today we've got a really exciting guest with us. Why don't you introduce her? Yes. I've been waiting for this podcast, specifically just for OCP. But today we actually have Jenna Bowler-Shoring from OCEAN. And she and the Veit Harthen and I'm working for OCEAN. So we just had a great panel with her this morning. We're excited to dive in a little bit deeper into what OCEAN is up to. Yeah, thank you, Denise and Allison. I'm really happy to be here with you today, and I appreciate you inviting me onto the show. So Jenna, why don't we just start with a bit of background on OCEAN and your background and how

you came to be the Vice President of Marketing. Sure, I'd be happy to share that. So at OCEAN, we are pioneering a new data warehousing technology that optimizes for maximum efficiency for always on very compute-intensive data and AI workloads. And our unified data platform technology enables customers to consolidate various capabilities that would typically be in disparate systems into a single efficient platform. So by eliminating the need to move data in between systems and by bringing more capabilities directly to their data in one platform, our customers typically realize a 50 to 90% savings in the cost, the system footprint,

and the energy consumption of their compute-intensive workloads. So with Drewby to OCN, I actually was working at Cthulhu Valit 2 prior to that, but I had worked with our CEO, Chris Flatwin, at a past company where they had revolutionized the data storage industry. It's a company called CleverSafe, and he was one of my favorite executives to work with in past jobs.

So when I heard about the mission and vision of OCEAN, I had to hop on board, and I was marketing employee number one. Fast forward three years, we've got a team of seven. We do global marketing and communications for OCEAN. And we're just hitting our growth phase right now. So it's a really exciting time. Nice.

It's a very exciting time. And OCEAN is really known for delivering real-time warehouse to large volume of data sets, right? You share a little bit more about the technology you're delivering and how you're bringing that to market. Absolutely. In terms of leading to mine data really fast, the type of software architecture that our team has pioneered has actually gone against the curve when you think of how the data warehousing market has evolved. So we bring our compute adjacent to storage. We've partnered closely with Solidigm to make sure that we're able to do that on an all-NVMe SSD architecture.

So we leverage the most modern hardware and compute cores and things like that, and then write our software layer to maximize all of the parallelism and the efficiency within that stack. So for our customers, they're able to really boost their performance without having to waste their spend, manage data faster, process data, ingest data faster, run machine learning directly on the data, prepare data, explore data, and do more, like I said, from one platform than they can with other types of solutions. We're really focused on maximizing that efficiency benefit, having a really tight data layer at the foundation, doing as much as we can at the foundational layer, and then bringing more capabilities, like I said, to the data.

So customers are having to move data, which obviously introduces more points of vulnerability, more security issues. More teams have to manage those data pipelines. So we're all about being as efficient as possible. Now we're at OCP, land of the hyperscale, with over 50% of the world's data center capacity expected in hyperscale environments by 2026. How do you see data warehousing changing? It's a great question.

So I think to date, the way that the data warehousing market has evolved has been around convenience, elasticity, being able to spin up new environments. But what we're starting to see with customers that operate on a very always-on basis is that the cost quickly gets out of control and actually is very unpredictable. It's not uncommon for me to talk to customers who are aggressively deleting data. They're trying to introduce constraints on their environment so that they can manage that cost. So I think we're going to see more of a backlash around inefficient usage, lack of predictability around spend for data because as customers need to do more, particularly in the age of AI, they're going to need to be more efficient in that core foundational layer. And they're not going to be able to afford to just waste money at kind of data preparation and data analysis layer. We're seeing a lot of vendors trying to tackle

data warehousing for AI. What is OCE doing differently? Great question. So I think a lot of software companies at the moment are putting AIs in their messaging, but not always clear what the value prop is. For OCEAN, we address AI in a couple ways. So in one is in that kind of 80% of an AI data pipeline, that's the data preparation stage, data exploration, just getting your data ready for AIs. That's an area where we help customers, again, trying to reduce the amount of complexity in just getting data ready to be put into a predictive AI or a Gen AI pipeline, enabling customers to do that more efficiently with fewer resources, fewer teams, less cost is something that OCEAN tackles today for our customers. And then we also have in-database machine learning capabilities. So you launch Watchgear, OCEAN to ML. It brings ML directly to data in OCEAN.

So if customers want to deploy and train a model, you have a variety of models, a very large catalog. They can do so directly in OCEAN or they can do some level of exploration, preparation, and send data to another system. What's the customer response been to these

solutions you're talking about? Great question. So, so far, our customers are pretty happy, which is great for us and something we prioritize. You know, as a newer provider to the market, we don't take any customer for granted. And so, we actually have an entire practice that we call our customer solutions and workload services team, and they partner closely with our customers. You know, at the level of computing that a NoSignt customer is doing, it's not an easy task to migrate to a new platform or station. And so our customer solutions team really partners with our customers to understand the data pipelines going into their solution, how they need to use

data, where there are opportunities to be more efficient with the processing of their data, or even the pre-processing of their data. And then they actually show our customers their working solution in a production-like environment before they sign that final purchase agreement, which sounds ridiculous. I know, why would we do all of that work up front before customers even signed with us? But that's what we believe is going to set our customers up for success. So by the time they know what they're seeing, they've already seen everything working. They're realizing value from day one.

And that drives incredible stickiness with our customers. Sure. When I came back and partnered on the next project, they want to grow. They want to expand their systems. That's kind of our go-to-market strategy now and something that makes us different. Yeah, this morning you talked a lot about sustainability and you seem to have such a passion for it, which I think is so great. But how do you at OCEAN really approach performance and efficiency within the solutions you offer? And how do you with your customers evaluate against your proposition? Sure. So for performance and efficiency, that was really the founding principle of OCEAN, according

to reasons why we partnered with Solidigm to have an all SSD architecture. No HDDs. We cannot run on HDDs. We only run exclusively on SSDs because ultimately we were originally focused on price performance. So maximizing the performance of a solution at the lowest cost using the fewest amount of resources and industry standard hardware for cloud instances. And what that actually translates to now is an incredible energy efficiency benefit. We're really focused on usage-based pricing models, on helping you spin up and spin down boost sources in a more elastic environment, working the cold storage tiers and dragging the network before you can get it into the compute tier.

From our first day, we've always optimized for performance at scale and for efficiency. So now we can show our customers kind of a footprint of maybe their existing solution or competitive solution they're looking at versus the footprint of an OCEAN solution. And it's a drastic comparison. It's often, like I said, a 50 to 90% savings going with OCEAN, which when our customers are thinking for the next three to five years,

it's an incredible savings for them. What have you seen at the OCP Summit that gives you better insight into data requirements ahead? I think there's a lot of data that needs to still be collected around energy usage, where the opportunity is to reduce emissions, energy, things like that. One of the things I talked about this morning is a lack of transparency on the software side. We see a lot of conversation around data centers and storage providers and cooling. But what about the software architectures

that were built for legacy hardware or the systems of five to 10 years ago? What I'd like to see is more transparency around the energy cost, running very intensive applications, especially AI applications, and more innovation around providing features and control to customers and users

around when they run their applications or how to optimize their applications with energy in mind. And one thing that I've been really happy to see here is so much collaboration. I'm touching my first time at the OCP Summit and I definitely think I'll be coming back. Nice.

We would love to have you back here and see you back here. But before we do that and jump into next year, why don't you talk a little bit more about the SSDs that you deploy in your solution? Just a moment ago, you talked about you could not really deploy hard disk drives, right? SSDs are a thing. But can you tell us a little bit about how you're using SolidInet SSDs, and are there any specific examples that you might be able to share? Yeah, certainly. One of the things that we've pioneered for our software foundation is what we call compute adjacent storage. It's our compute adjacent storage

architecture. Short name for that is CASA. And we named it that way because it really is the home of where a lot of our data processing happens. And so we leverage solid-idon SSDs directly attached into our compute cores, super high-core cal processors, to make sure that there's really minimal movement of data. The compute layer is right there where the data resides. And as much of the processing as we can, we push down into the storage layer.

So we don't even need to move data into that compute layer to do all of our processing. In customer examples, you've been able to show them how they move from, let's say, a legacy Hadoop type of architecture or a cloud data warehouse or database that's leveraging more legacy hardware. We can show them the entire stack from the OCEAN software layer, the Solidigm layer, the compute cores, how the entire solution would be more efficient for them. And we've been able to show them that 50 to 90% reduction, which when you show them that you're going from 10 racks to three or 12 racks to five, how that really is going to

impact the way that they manage their solutions. And one of the things I like to say, too, is that OCG would not exist without partners like Solidigm because it was really your innovation that enabled us to build this software on top to deliver this benefit to Hesrus. That's so cool. I'm so glad that we got some time with you. It was so much fun having you on our panel earlier today. And it was great to learn a little bit more

about OCEAN in this conversation. One final question for you, Jenna. Where can folks engage with you and learn more about the solutions we talked about today? I'm so glad you asked. So folks can visit our website, ocient.com, O-C-I-E-N-T.com. And they can also connect with us on LinkedIn. Pretty active on LinkedIn. Either way, you can get in contact with us. So yeah, I hope if there's anyone out there that needs a more efficient, compute-intensive solution for their data and AI workloads, please check us out.

We'd love to tackle your top data challenges. Awesome. Thank you so much for being on the show today. Thank you. Thank you, Jenna. Thanks for joining the Tech Arena. Subscribe and engage at our website, thetecharena.net. All content is copyright by The Tech Arena.