NetApp Intelligent Data Infrastructure Video

In Conversation: NetApp’s Intelligent Data Infrastructure

Earlier this summer I had the opportunity to talk with NetApp CMO Gabie Boko and NetApp VP of Product Marketing Jeff Baxter in a wide-ranging conversation about the power of data infrastructure, which extends well beyond simple enterprise storage.

As NetApp Insight, its premier customer event, kicks off this week in Las Vegas, I thought we’d revisit the discussion, as it provides nice context for what the company is expected to announce this week.

This video, sponsored by NetApp, streamed at LinkedIn Live in June 2024. The below transcript has been very lightly edited for clarity.


Steve McDowell: I am Steve McDowell, Chief Analyst at NAND Research, and I’m at NetApp with Gabie Boko, the Chief Marketing Officer, and Jeff Baxter, Vice President of Product Marketing.

To set the stage, the industry is in a unique moment where data has become business critical. Data can impact the competitiveness of the enterprise, but at the same time, it’s never been more complex to store and manage your data. We have hybrid-cloud, we have the rise of Edge. AI is touching everything. We’re under constant threat by bad actors. It’s a very interesting time for enterprise IT.

With all of that said, Gabie, how is NetApp addressing the evolving needs of the enterprise customers with your data strategy or your storage strategy?

Gabie Boko: Thanks, Steve. Really glad to be here with you. I think that when we think about NetApp all up, we are really leaning into what we call intelligent data infrastructure as a way for us to think about not just what our customers ask of us in the storage world, how to think about their storage, but also their evolving needs of data.

When you think about, like you already just said, increasing complexity, and what we’re really interested in is making sure that that definition becomes a new infrastructure paradigm, essentially, where we’re saying that data infrastructure is built on world-class storage.

It has to have a built-in single OS, has to have a great control plane, all of those things. World-class security has to be part of what you’re considering. So as we think about it, we feel like the needs have always been the same, but they continue to grow in complexity because data becomes more important and grows in complexity. We’re trying to evolve along with that message.

Steve: A big part of that complexity is the rise of hybrid multi-cloud. I don’t think there’s an enterprise out there that doesn’t have some cloud presence, sometimes more than one. Just look across the storage industry, NetApp has perhaps one of the strongest offerings of any storage provider in the public cloud. Why is that important to you and your customers?

Gabie: Yeah, that’s such a great question. We like to look at intelligent data infrastructure as one of two things, right? You can either be cloud-embedded, to really protect and secure and organize your data, or you could be cloud-conflicted, where you cannot really be in partnership or you’re kind of loosely aligned, or you might have a multiple additional operating system.

When we think and we have embraced our strategy of being embedded with the world’s largest clouds, we understood instinctively that most of our customers’ data is unstructured. Most of that is going to live in a hybrid, multi-cloud environment, and that the best way for us to be partnered with our customers is to be where our customer’s data was.

In our world, whether it be AI, whether it be security, whether it be just what you’re trying to manage in terms of your data from an everyday motion, we want to be where you are going to manage it, and so for us, that’s being in the three biggest clouds for sure.

Steve: One of the strengths of NetApp’s approach is managing cloud resources with on-prem with a unified control plane. You also have a variety of flexible consumption models and service offerings to make it easy to pick and choose my storage infrastructure. How does all that kind of fit together? How do you look at hybrid cloud as a business?

Gabie: I’m going to put it back to intelligent data infrastructure, and then I’m certainly happy to have Jeff unpack all of the great product perspectives in there. But when I think about intelligent data infrastructure, we’re really looking at the building blocks of how you’re going to organize and manage your data. And what we like about that perspective is that it is enabled by a platform that is strong and consistent.

It is also key building blocks of modularity, so you can do what you need to do when you need to do it. Pick your storage system, pick your services, pick your operation services and pick your partners. That’s really the way we think about it.

Bolstering that in my mind is a really strong operating system. You can’t do this if you’re not thinking about built-in from the ground up. All of our storage systems, obviously, totally embedded and enabled by ONTAP, the industry-leading data management system.

I think that when you really look at that, that is where a lot of our intelligence comes from as well. It’s the ability to control security, to control the services we deploy, and is different from those in the industry because that makes intelligent data infrastructure, again, choice-based, right? Need-based, not buy it all, and then upcharge for each additional thing.

On top of that though, if I had to bookend, it is a really great control plane, and that’s what we have with BlueXP. When you think about BlueXP, you’re really thinking about not just a control plane that you have to use, you must use. But if you want to use the AWS control plane or the Google Control plane or something else, then you can absolutely do that because of the rich depth of how we construct it and how you can deploy those services.

Last but not least, when you get to all of those, what you’re talking about is not just choice in terms of what you pick or how you deploy it, but you’re also deciding how you want to pay for it. And those are consumption models, right? That’s through Keystone, right? Do I want to buy this through a SaaS model? Yes, you can do that.

So when I think of IDI, I’m thinking intelligent data infrastructure is modular, it’s custom, but not custom. I would say maybe personal. What are your data needs today and where are my data needs going to be? I want a company who has thought about all those pieces, connected the dots, built that functionality and consistency in, so no matter what I have, I know that it can grow with me and I can swap out or add as I move.

Steve: So Jeff, are there unique capabilities that NetApp brings to hybrid cloud?

Jeff Baxter: The ability to really have that same unified data plane across all the different clouds, as well as on-prem storage, brings us the ability, it gives us basically the surface area to address any data challenge.

With all of these built-in capabilities, Gabie mentioned some of them, we’re able to bring to bear across the hybrid multi-cloud in a very common and differentiated way. So when you get storage efficiency or when you get security or when you get the ability to move data around, all of that is done through a common platform, all of that is done via common API, so it really simplifies things for our customers. I always talk about as learn once, do many. Regardless of where your data needs to be today and tomorrow, you’re going to have a common approach that works across the hybrid cloud.

Steve: Apart from hybrid cloud, when you talk to enterprise CIOs and enterprise leadership, period, top of mind is cybersecurity, data’s business-critical. I need to protect those tools. And this is one of the areas in which NetApp I think stands out in the industry. You have perhaps one of the most compelling cyber protection stories where you do real-time ransomware detection and things like that, or malware detection.

Talk to me about how you look at that and how you ensure the safety and reliability of the data that we’re putting into your infrastructure.

Jeff: I’ll say NetApp tries to be very pragmatic. We make claims, but we do make a very bold claim that we’re the most secure enterprise storage on the planet, and we have that backed up through numerous testing certification. We’re actually the only enterprise storage that’s validated by the NSA to store top secret data.

That tells you that although not everyone’s running around with top secret data, it’s a sign of trust. And we’ve spent the last three decades really building security into the foundation of everything we do. Now, as we’ve extended out to the cloud, we’ve really built an entire framework that allows us to embed security across every step from sort of inception of your data to protection of your data to responding to incidents like ransomware.

We can handle basically the whole NIST framework if you look from all the way from identifying your data with things like our BlueXP classification service from protecting your data with built-in snapshots, encryption, from being able to detect ransomware in real time.

As you mentioned, our autonomous ransomware and protection that’s now being enhanced with AI models to be able to detect with greater than 99% accuracy ransomware in real time when it’s attacking your system, and then be able to allow you to respond to it through our unified control plane and to be able to recover your data back in minutes, all of that is built in all of those included in the NetApp solution for cybersecurity.

Steve: As I understand it, NetApp is the only mainstream storage provider that’s doing real-time, malware detection in the box with that level of integration.

Jeff: Yes, it’s built in, it’s integrated, it’s available on the box. It’s also available in our cloud storage offerings as well in many of our cloud storage offerings. That’s unique and differentiated from NetApp as well.

The idea of bringing AI ML with a constantly updated model and building it directly into the system, because we know ransomware is always evolving, it’s evolving on a daily or even hourly basis, and ransomware increasingly is being designed with the help of AI, and it’s one of those, the only way to fight AI-designed ransomware is going to be with AI-empowered ransomware defense.

Steve: Does that put stress on the system? Does this impact my performance?

Jeff: No.

Steve: Do I sacrifice anything?

Jeff: You know, that’s a great question. One of the beauties of ONTAP is the way it’s written, it is just written incredibly optimized for flash, and being able to use that within the system allows us to optimize. Since we’re already in the data path, checking for malware is as simple as any of the other checks we do for metadata.

We’re really looking at file patterns, file signatures. So the impact of that is at most measured in sort of the single digit percentage points. It’s just kind of baked into the performance of our overall system, just like all of our performance testing and everything is done with storage efficiency turned on, it’s done with all of our other data management features turned on.

We think ransomware detection is just a basic inherent data protection feature that should be included in your enterprise system, and it shouldn’t come in any significant performance penalty.

Steve: I’ll switch gears a little bit and talk about sustainability. Every enterprise has sustainability goals that they’re meeting. In some geographies, it’s regulated by law, and when you look within an enterprise, IT is disproportionately responsible for a lot of the issues.

How do you look at this? And NetApp talks a lot about power efficiency. How does that play into the sustainability story?

Jeff: There are a couple ways we look at sustainability. NetApp takes sustainability very seriously as a company. We were recognized by EcoVadis with a gold rating just around our general ESG sort of initiative. So that’s what we do as a company. We take sustainability incredibly important in terms of our own carbon footprint and everything like that.

Then for customers, I think there’s really two broad areas we help them address. One is to understand what their sustainability footprint looks like. We have built-in APIs into all of our products that will tell you about your power utilization, everything you’re using, and then as part of our BlueXP control plane, we actually have a sustainability dashboard that’s just provided at no cost to all of our customers that will help them rationalize all the inputs from their entire NetApp install base into talking about this is your carbon footprint.

We also work with things like the MIT PAIA consortium to be able to do lifetime carbon assessments of all of our new storage systems that are coming out, so that when you’re buying a storage system up front, this is what the carbon footprint of that system is, and then we build that directly into our APIs to also report it back to you programmatically. That’s really the first part in terms of helping customers understand their footprint.

Second is really helping them improve their footprint. We pioneered storage efficiency back now several decades ago. We’ve continued to enhance our storage efficiency because obviously the greenest storage is storage you don’t use. So if we can enable you to have the smallest storage footprint, that’s going to help.

Then we’ve done just incremental improvements basically with every release and every enhancement. So our newest systems use titanium-rated power supplies to improve their power utilization ratings. We continue to improve our storage efficiency.

In the latest systems we announced, our storage efficiency numbers improved because we’re able to use offloads within the system to improve that overall efficiency. So all of that combined with greater and greater higher-density flash to just reduce the overall rack unit footprint just reduces our environmental footprint in a significant way on behalf of our customers.

Steve: And flash itself has an intrinsic sustainability story over spinning hard drives.

Jeff: Yeah, flash, both from an individual flash drive on a terabyte-to-terabyte basis is going to be more power-efficient. And then the pure density of them. We have a comparison. If you have customers with 1.8 terabytes SaaS 10K drives, which are very, at this point, legacy, but still a massive install base out there, if you move from that to the newer 15-terabyte or larger-storage drives, you can put two and a half racks worth of storage into a single sort of for-you space. So just if you can conceptualize both the space savings as well as the power, carbon, cooling savings, it’s a mess.

Steve: There’s a perception among many that flash storage is for high-performance workloads. That’s shifting, right?

Jeff: That’s absolutely shifting.

Steve: With QLC-based capacity-optimized storage arrays, we can now push flash storage, all flash storage down into the near-line storage tier.  

Jeff: I think that the first decade of flash’s life, really, people thought about hybrid flash as the baseline, and all-flash was the exception, all-flash was the exception for high performance. Now, we really think of all flash as the baseline and hybrid flash is the exception for truly secondary or archival workloads, right?

NetApp is unique in that we still do offer hybrid flash. We have a thriving hybrid flash business that runs on the exact same storage operating system, NetApp ONTAP, so it’s all unified environment.

Increasingly, our customers are adopting that when they want to tier cold data off of an all-flash system, when they want to back it up. When they want to use it for something archival that’s never going to be used again, most likely, that’s a great place to still use disk drives, but everything else has moved to flash.

With the introduction of capacity flash, we’ve said our NetApp AFFC series that we introduced over the last 18 months or so has been the fastest growing product in NetApp history, specifically because, as it turns out, customers really like it. If you can get a high-performance flash system at a lower cost, it’s a great value proposition. And so we’ve seen just tremendous uptake at capacity flash.

Steve: Talk to me about NetApp’s legacy, ONTAP, it’s over 20 years old. Is it a struggle to keep it fresh? Is it showing its age?

Jeff: No. So 1992 I guess was when… I mean, and it existed a little bit even before that. But this is not your grandfather’s ONTAP, right? It’s kind of like the ship of Theseus, right? If you replace a plank one at a time, is it still the same ship by the time you get to the end of it?

ONTAP has gone through a couple major rewrites during that period of time when we modernized it and really made it… And the modern ONTAP is very modular, very easy for us to upgrade individual pieces of it built, on a strong foundation. The last major update of it that we did, we enabled all of the scale functionalities and we really built it to be optimized for flash, and that has carried us forward.

ONTAP is able to stretch into all sorts of new workloads now, and AI obviously being a key one where it’s just shown its maturity, but also its power and performance. And I think we’re blessed to have it as a foundation, but every day we’ve got thousands of developers continuing to enhance it. It certainly continues to be a moving target of innovation.

Steve: When I look at NetApp, to me that’s one of the strengths is your legacy. You’ve been doing this… You almost invented this industry in the 1990s, and you’ve continued to evolve and stay a step ahead of where enterprises are landing. Where are we going next?

Gabie: Ooh, where are we going next? I would say you’re going to see us lean into more platform type of functionality. Really, that’s that more built in. Where can we provide more services? Where can we provide more usability and experience functionality? But I also think what we’re really going to lean into is what I call maybe more personal workloads if I think about it, it’s really about what are the workloads that are going to work for you and how do you maybe drop a workload in that’s already pre-configured? Not just a solution, but I want to deploy sustainability in my area. This is the KPIs that we know run, this is how I’m going to deploy that workload. Same with ai. I want to go play a little AI in the NetApp sandbox. I’m going to go do that. So I think you’re going to see us lean into that more and more.

Steve: That’s a great tease. We’ll be hearing more about this at NetApp Insight in September, October?

Gabie:

September 23rd through 25th in Vegas.

Steve: Excellent. Let’s talk about AI for a second. Does storage need to change to support the trajectory that AI is on in the enterprise, or do I have to do anything different as a storage administrator?

Jeff: I think it depends on what your storage is, and I’m a little biased, obviously. I think what AI is, and the key thing to me about AI is AI is not one workload, right? AI is at least five different steps along a data pipeline, if not more. And with some traditional, I think, legacy storage or flexible storage, you’re basically creating a storage silo for each step of that process.

We believe that if you have a good, unified data storage solution that has these built-in data services, that you can use a common framework for just about all of your AI training workloads. So it certainly places more demands on a system. And when people think of AI placing more demands on a system, they typically are thinking performance.

And that’s certainly true when you’re talking about high-end model training. High-end model training can stress anyone’s storage, and we think we have absolutely high-performance systems that we validated with NVIDIA to be able to support both, for example, NVIDIA BasePOD and NVIDIA SuperPOD configurations.

We can stand up on the performance frame as well as anyone, but there’s so much more, in terms of doing inferencing, retrieval-augmented generation, fine-tuning. All of that demands more data services and it demands storage that can act as more than just a single weapon. It needs to be a Swiss army knife of AI. And I think that’s really what NetApp is able to bring to the table.

Steve: That’s where I think AI is going to land in the enterprise. Enterprises are not training their models by and large. They-

Jeff: No. It’s going to be a handful of companies that are going to create the really massive models that we see out there and keep them up to date. Just because the number of GPUs required and the number of data scientists and the amount of data necessary to train them on is exorbitant. It is well beyond the imagination of most enterprise IT.

It doesn’t mean they won’t do some smaller-model training of lower-parameter models for specific workloads, and that’s certainly something that can be supported in enterprise IT, but far more likely is they’re going to take these large models, whether large language models or other sort of models and do fine-tuning on them.

Or increasingly, where we’re seeing a ton of interest is in retrieval augmented generation, taking a large language model and adding vector embeddings with your own private data on-prem in real time so that you can add the intelligence of your own data into one of these large language models and really get the best of both worlds.

Steve: NetApp has a powerful story. We’ve talked about a lot today between the two of you, but Gabie, I’ll let you close this out, is how is the NetApp story resonating with your customers right across Enterprise?

Gabie: It was, it’s definitely a journey on how to speak to a new narrative. We did a lot of testing and a lot of conversations both with industry analysts like yourself, with our partners who are absolutely critical to our ecosystem, part of our story, because we believe in built-in means, including our partners and infrastructure, but also most importantly our customers, and our customers felt like maybe we lost a little bit of our way. We got a little bit too far ahead of them.

So right now, when they come back and tell us our three biggest priorities, not in any order, are obviously “Please help us connect our data across the clouds. Please help us secure that data. Please help us really understand and evaluate how to see that through a single control plane.” When we think about that, we are thanked for really… “Thank you for a new narrative, but also thanked for… Please thank you for including us in whatever your strategy is, and the development of your future roadmap.”

So right now it feels like we’re happy with that and they’re happy with that. I think with any brand though, it’s about evolution.

And it’s about listening, and that’s what we want to continue to do for our customers. So as their needs with data evolve, our narrative, our strategy will evolve. That’s what I’m most excited about with Intelligent data infrastructure.

Steve: Wow. What a great note to end on. Thank you, Gabie and Jeff, for taking the time, and thank you for tuning in. For more information about NetApp and what they’re doing with their intelligent data infrastructure, it’s netapp.com.

Gabie: Netapp.com.

Steve: It’s about that simple. Thank youi

Disclosure: The author is an industry analyst, and NAND Research an industry analyst firm, that engages in, or has engaged in, research, analysis, and advisory services with many technology companies, which may include those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.

Leave a Reply

Your email address will not be published. Required fields are marked *