NYSE Media Week CXO Series, theCUBE + NYSE Wired | Doug Gourlay, Qumulo

Clips
More from NYSE Media Week CXO Series, theCUBE + NYSE Wired

Douglas Gourlay

CEO

Qumulo

play_circle_outline Growth of unstructured data, particularly in video and AI applications

play_circle_outline Focus on culture, customer success, and high-quality product at Qumulo

play_circle_outline Unlocking Data Management in Multiple Clouds: Qumulo's Standout Run Anywhere Model and Global Data Fabric

Info
Transcript

Doug Gourlay, Qumulo

Douglas Gourlay

CEO Qumulo

Qumulo CEO Doug Gourlay discusses the company's focus on managing vast amounts of data, especially in AI. The need for flexible data storage solutions across different cloud environments is crucial. Unstructured data like video is key for AI and machine learning. Qumulo's "super cloud" approach offers a consistent experience on various platforms. Gourlay stresses company culture and customer focus. Qumulo sets itself apart with a versatile data management model and global implementation. The goal is to keep innovating to meet industry demands.

explore Keep Exploring

What are some factors contributing to the rapid growth of unstructured data, specifically in the context of video technology advancements and data retention challenges? add

What are some impressive aspects of the organization mentioned in the text? add

What are the four core things that set our organization apart from others? add

bolt Powered by CUBE AI

Doug Gourlay, Qumulo

search

>> Hi, everybody. Welcome back to our CXO Series in collaboration with NYSE Wired. You're watching theCUBE. I'm Dave Vellante. Really pleased to have Doug Gourlay here, the newly-minted CEO of Qumulo. Doug, great to see you in our studio here at NYSE. This is awesome, right?

>> It's a fabulous being here. I like the energy. The fact that, seeing OpenAI and everything going on here, it's a great day .

>> Yeah, Oracle rang the bell today. Sonny Singh's coming in-

>> They've never done that before, have they?

>> Yeah, right. Oracle's smoking,

>> Killing it.

>> As your former firm, wow. But tell me, why Qumulo? Why did you choose this company?

>> Really simple, Dave. If you look at the amount of data that's being generated today, it's unparalleled how much is being created both by generative AI, as was phones, surveillance systems, cameras. More data is being generated now and will be generated in the next three years than has been created in the history of humanity. If you look at that, you go, wait, it's being generated and it has to be retained. You combine that, the amount of data is exceeding Moore's Law's ability to absorbing, which means I have to build new data centers, go to hosting facilities or go to the cloud, or some combination of the three. And what Qumulo was uniquely positioned to do was create a cloud data platform that worked in the public cloud, the hybrid cloud, and the private cloud, allowing customers' data to be effectively managed wherever it's located. And that was a very unique value proposition, that looking at the needs of AI to consume data really spoke to a tremendous opportunity.

>> It's true. In fact, we coined this term, super cloud. It was sort of tongue-in-cheek-

>> Yeah, no, you're right again.

>> But it's going beyond multi-cloud, so you have a consistent experience across cloud. We always use Qumulo as an early example of super cloud. Now many have-

>> You can't say Meta cloud for other reasons, it's just a little too obvious.

>> That's right. But some people do use that term, Meta cloud. Okay. So the fundamentals are great. In fact, we're blowing away Moore's Law. I mean, if you just look at what's happening inside of these Apple machines, let's say Moore's Law's 40% a year, we're talking about well over a 100% a year in terms of performance improvements. Now with AI and what Jensen's doing, we're talking a 1000X in eight years versus a 100X in 10 years. So those fundamentals are good.

>> It demands unstructured data, the primary vehicle for fueling learning, training, inference, tuning workloads.

>> Okay, I got to ask you. So ever since I've been in this business, unstructured data has been 80% of the data. It's got to be 99.9% of the data by today.

>> It's probably north of 90. And I don't want to give a hard statistic on it because your mileage may vary, but absolutely, yes.

>> But there's no question. Unstructured data is growing faster than transaction data, and there's much more of it in video has changed the game.

>> Right. Well, you look at your phone and as you pointed out, 2007, didn't do video. First iPhone that started doing video like VGA 640 by 480. Today I've got 4K HDR stereoscopic, so I can play it on my Apple virtual headset, which I love playing with. The amount of data that can be generated off one cell phone is orders of magnitude more than could be done even five years before. And then we save it forever. If you look at video surveillance data alone, you're looking at data where the retention requirements are sometimes aligned with the statute of limitations on prosecutorial conduct. It's like, do I want to throw away the data that could exonerate somebody or throw away the data that could prosecute somebody, or do I need to save that for a very long time? And that's what starts creating even more impetus and more data that's out there. There's further fueling AI systems, image recognition systems, video recognition systems, just a tremendous amount of content.

>> Now, am I right, this is your first CEO stint?

>> Yes.

>> So what did you learn from Jayshree that you want to bring to Qumulo? And what fingerprints do you want to put on Qumulo that are maybe unique to Doug Gourlay?

>> It's funny, working for Jayshree was one of the greatest experiences in my life.

>> I bet.

>> Her energy level, her passion, her ability to spot an emerging market trend and redirect abject amount of resources to focus on. It was unparalleled. And when I look at us as a company, that energy level, that focus on the customer, is something that was always very natural to me. But now seeing myself in this seat, there are days I look at myself in the mirror and go, oh my God, am I really here? What am I doing? And as much as I sometimes don't admit it, there are multiple times a day, I say, what would Jayshree do? And what does it come down to? Client conversations, being aware of the market, building a tremendous network of technology peers. These are all things she's done that were amazing and helped catapult Arista into the stratosphere. And being able to learn from that first-handed, and then take that and apply it. I'm already seeing results in my own organization from applying some of these basic blocking and tackling things that she did that were so valuable when applied into a new company.

>> Culture's an interesting thing-

>> It's huge....

>> for a CEO. Right? And when you first, when you're a kid, you hear culture, you think maybe it's touchy-feely-

>> It's like yoga, what are we talking about?

>> Right? But is really, begin to understand how critical it is. How would you describe the current culture? A lot of times culture is just there. It's bubbling and it just needs to be molded. And your customer-first inclination is interesting and obviously effective these days. But how do you see it at Qumulo? Is it there? Is it bubbling and you're shaping it, or was it already there or is it a big pivot? How do you see that?

>> There were some parts at organization that were incredible, the focus on customer success, the foundational quality in the product, the fact that the company has over a 1000 customers in production today, not, with no customer concentration, it is life sciences and healthcare, it is media entertainment. You can't find a movie or TV show or animated character made in the last five to 10 years that doesn't run on and exist and stored on Qumulo in some way, shape or form, whether through a VFX shop or primary production facility. We have customers today that, we are storing on Qumulo the cure for cancer. Now we need to find it within our data. And another one say, we're generating 750 terabytes a week of data that in it is secure for ALS and other degenerative neural diseases. Now we have to build the models to unpack it and figure out what protein interactions or neuron interactions are causing this. I'm like the level of responsibility you have for data preservation in that environment, unparalleled. The closeness you get to the customer's workflow and use case, . And so you take that, and have a company with a foundation of customers, huge customers, customer success, high quality, what an incredible foundation. Yet, there's some areas to improve, always. Where it's things like the training and education of our go to market organization, the narrowing of the focus into the use cases where we have the strongest competitive advantage and differentiation, the evolution of our technology stack to maintain pace with the market changes that require that, both in performance but also features and capabilities, and expanding that cloud-native portfolio to the broadest cloud-native storage portfolio in the industry to adapt.

>> So if you think about the storage business, and it's changed quite dramatically when you had mainframes, then you had EMC, became the sort of dominant model, and the EMC and NetApp were the two sort of independent companies. And then you had some guys that tried to escape, get to escape velocity, not many, 3Par kind of got there and then got acquired and Isilon and on and on and Compellent, et cetera, et cetera, et cetera. And then you had Pure came out and-

>> Held up pretty well.

>> Doing pretty well, we like that. We like independent companies. Obviously EMC's now been absorbed. Storage companies have evolved from boxes essentially, to stewards of data that are software companies. That's a different business. It's different marginal economics-

>> We have no hardware.

>> How do you see it?

>> We have no hardware. We have a run anywhere model. We are in production today on Dow. We're in production on HP, we're in production on Cisco. We're in production of Supermicro, we're in production of Arrow. We're in production of Fujitsu and . This gives our customers unprecedented freedom of choice. By the way, we're in production on Amazon, Azure and Google as well. So when we say run anywhere, it's run anywhere. Same software, same management, any environment. That solves supply chain issues, it allows better unit economics. It allows you to deal with vagaries of market, commodity pricing changes in much more flexible matter for our clientele. Now, the other thing it means is that our clients don't have to go re-certify our net rule appliance every time we rev a NIC or rev a software version or hardware version. Oftentimes we're running on the same hardware platform, same Linux version with the exact same tooling the customer's already tested and qualified, so the same software lifecycle management, same image management, same vulnerability management, same CMDB, same ITSM. All of that consistent massively buys down the customer's time to getting us into production and time to value. And in today's environment, like in life, the only resource we have we can never get more of is time. If you have a two-year lifecycle on that GPU from the time it ships to the time it's obsolete and the next one comes out, do you want to waste a year of it testing and qualifying and building the infrastructure? Or do you want to be in production in a month maximizing that time to value, maximizing that two-year sprint you have to gain that competitive advantage and crush your competition?

>> So in my experience, what you're saying is a good idea, right? The singular experience across any estate, cloud, on-prem, any cloud, et cetera. A lot of people are doing that now. A lot of people think, oh, I can do that. And you know what it's like in tech. Oh, that's easy to do. And then you start doing it and it's like-

>> It always looks easy from a distance, doesn't it?...

>> well, this doesn't really work. And there's this thing called recovery and on and on and on. And integration and all the systems you have to be part of, these-

>> And even partnerships, yes-

>> take a lot of effort. So my question is what is your sort of differentiation from the cartel and the upstarts? Is it that experience that you have and that singular cross estate platform? What do you see it as your differentiators?

>> Four core things, Dave. The first is believe strongly we're the only organization with a true run anywhere model. And that is not just in the cloud, it's x86 and it's R. Then cloud-native. We are bringing the same unit economics as the on-premises environment, which is the most competitive unit economics, to the cloud. Our customers are seeing almost identical TCO in the cloud versus on-prem, especially when you factor network infrastructure, power, facilities, depreciation schedules, staffing, very complex environment, very turnkey in the . Then we have the global value space implementation, building a global data fabric to connect the data that's spread all around the world. Simple example, one of our customers, a large research facility, seven research centers, 350 petabytes of data, two data centers, three cloud providers, and now a new AI processing facility in Texas. How do they get the data from the research facilities and the clouds, we're talking 100s of petabytes and get all of that to the AI processing facility, do its job, and then bring that data back? Building this global data fabric to connect all of those? Incredibly powerful. Or think of it in media and entertainment right here in the studio. How do you have a follow-the-sun edit suite, where I start off editing in London and then New York and Vancouver and LA, finish up in Hong Kong, then Mumbai and rinse and repeat. Can we have a single file stored in one master version that is modified by the 10 different developers concurrently as it moves around the sun or as we move around the sun? Absolutely.

>> So it's not a copy model. Copy here, copy there, copy there.

>> Single instance. This is lock-based replica, strictly consistent. No data is lost and no one overwrites or steps on each other.

>> So this is interesting because I hope, don't hate me for saying this, but they call Qumulo the son of Isilon. But what was brilliant about Isilon is they were able to do that, but within a data center, right? They weren't able to do that across a global namespace-

>> Now we're doing it within a planet.

>> And that was the interesting thing about Qumulo when we first saw it. We said, okay, this is Uber. This is super cloud kind of approach.

>> Yes.

>> So what do you want to do with the company? You want to be here listed someday?

>> Who doesn't want to be here? Rang the bell once. I would love to do it again, Dave.

>> Yeah. Awesome. Doug Gourlay, thanks so much for coming on theCUBE.

>> Thank you for having me.

>> Really great to have you. All right, keep you right there for our next guest. This is theCUBE at NYSE, with NYSE Wired. I'm Dave Vellante. Be right back, right after this short break.