NetApp Cloud Building the Most Differentiated AI Era Storage Platforms

Good afternoon, everyone. My name is Punit Dewan, and I'm joined by Scion with me. We're part of the cloud storage product management organization.

And today we'll be talking about how NetApp is helping our customers take advantage of one of the most profound technical innovations and transformations we are going through in our lifetime. You know, cloud and AI are impacting almost every business, every day of our lives, and NetApp is very proud to be part of this transformation. Now, digital transformation thrives on data, digital transformations happens everywhere, whether it's on-prem, across Azure, Google Cloud, AWS, everywhere, and NetApp is everywhere, wherever our customers are, and that's what we are most excited about through the rest of the pres.

Through the rest of this presentation, we'll talk about, um, the NetApp portfolio across hybrid multi-cloud with specific focus on our partnerships with Google, Azure, and AWS. We'll talk about use cases and customers who are leveraging NetApp technology to fuel their business objectives. We'll also talk about deep dive into the portfolio, um, components across our one p, our CVO, as well as Insta cluster.

Uh, and then we will deep dive into some of the new announcements we have made at Insight. We will also talk about some of the innovation we are bringing with our partners, especially, you know, around ai. And then we'll wrap it up.

And if you have got any questions in the middle, please, you know, feel free to ask, like you were asking in the last session. You know, very, very good questions. Uh, let's start with, uh, the NetApp, um, cloud portfolio.

Uh, when I joined NetApp two and a half years ago, one of the key reasons I joined was NetApp was the only data infrastructure company that actually brought the hybrid multi-cloud data fabric to reality. That was two and a half years ago. It's still the case.

I'm pretty sure that's gonna be the case for a long time. That really excited me and that really excites our customers, whether their customers are on premises across any major cloud, sovereign cloud, whether they're running, you know, their data infrastructure or data platforms. NetApp is with them across their data journey.

And AI is really fueling, you know, the need for the data to be everywhere, to be utilized everywhere, to make decisions where the data and compute are located so that customers can get faster decisions, lower, you know, their overall cost, and drive real business outcomes. Uh, by bringing the power of data with ai. If you, if you look at our overall portfolio, uh, we have, uh, a deeply integrated portfolio of our first party offers with Azure, Google Cloud, and AWS.

Um, uh, this is for the customers who want that native experience within Hyperscaler. If I'm an AWS, uh, customer, I want the native AWS experience, you know, for my storage. And similarly across the clouds, our first party offers are deeply integrated into each hyperscaler native experience.

So customers get that benefit of ONTAP while still operating within each hyperscalers environment. There is also a set of customers who like the full flexibility of ontap, right? They want the same experience.

OnPrem, they want the same controls, the bells and whistles. They want to control the versions. They wanna self-manage ontap, right?

And for those customers, we have something called cloud volume on Tamp, CVO. Um, lot of hundreds of customers use CVO, you know, to manage their data infrastructure across hybrid multi-cloud environment in the way they have been doing that on-prem for decades now. And lastly, it's not just about storage, right?

Most of enterprises have the need to capture real time data. They have realtime data streaming pipelines. They need to analyze vast amounts of data, you know, with no SQL databases.

They want to run real time analytics on the data that's coming from all different kind of sources. Uh, how do you do that? You know, putting these data platforms together across different operating environments is challenging.

It adds cost, it adds complexity when things are not repeatable. Operating at scale is really hard, and that's where our Insta cluster portfolio comes in that provides our customers a hundred percent open source based data platforms to stream data to search, do intelligence search with the data to analyze vast amounts of data. So, and, and, and do, and, and doing that in a consistent fashion across hybrid multi-cloud, reduce overall operational burden, reduce overall cost, while not locking into any hyperscaler technology because it's a hundred percent open source compliant.

And our expansive portfolio has helped largest of largest enterprises, you know, trust NetApp to run their mission critical applications. We have more than 5,000 enterprise customers now leveraging our data and ins cluster platforms to fuel digital transformation. We have more than two exabytes of data under management now, up 50% year over year.

It's a massive growth, right? This is a great proof point of the trust that customers have placed in us, you know, to fuel their most, uh, mission critical applications. With Insta cluster, we have more than 20,000 nodes under management, across hybrid multi-cloud and, and, and, and NetApp's storage portfolio, as well as the data platform's portfolio.

We are going across different use cases, whether it's about traditional file sharing for users, whether it's about file sharing for high performance applications in H-P-C-E-D-A, whether it's about running databases like SAP, Oracle SQL Server, whether it's running about, um, um, uh, data transformation pipelines, whether it's about, um, bursting to cloud customers have trusted NetApp across a wide range of traditional IT, operational transactional databases, VMware environments by standardizing on one single unified storage solution. Only NetApp can do this. There are many, many examples for such customers.

You know, it'll take us ages to go through, you know, um, all these examples. But a few ones, you know, look at Infor, for example. Uh, they're using the CLO clover leaf application.

Uh, with NetApp. They needed to process tens of millions of small messages, one kilobyte, two kilobyte messages. They needed a high performance file system running in conjunction with, um, Amazon EKS with the right scalability, with the high performance, with the right availability, uh, across different AZs.

They standardize an ontap. Adidas, you know, when they were, um, um, thinking through how to bring SAP HANA on AWS with, while still meeting very stringent recovery time objective. They picked FSX for ONTAP to build their SAP HANA databases.

Customers like Scotiabank, you know, we talked about, uh, earlier, um, uh, today as well, HomeTrust, when they were moving their VMware environments to Azure or Google Cloud, you know, they really liked the simplicity of an ONTAP experience everywhere. The cost advantages of tiering bringing to, uh, VMware solution, you know, the rich R-P-O-R-T-O and data protection capabilities. We have built over decades, you know, for on-prem customers now available in the cloud as well.

Customers like tho routers, um, we were with them in their journey when they moved from on-premises to AWS as or less to Azure. One single consistent ONTAP experience across hybrid multi-cloud. Who else can do that?

And this is, this is the reason which really makes, you know, us, everyone at NetApp so excited about really simplifying these customers use cases, um, across different workloads, across different operating environments. Are you going to discuss at all on the open source aspect of, uh, Insta cluster? Like to the extent to what extent you have a community, you know, of, of, for development or otherwise maintain it as an, or support as an open source project?

So, Uh, Insta Cluster is, is actually the product offering name of the product. Yeah. A you know, I have to stand up.

Okay. Sorry about that. Uh, Insta cluster supports, um, you know, we, we provide productized versions of Kafka, Cassandra, and open search, you know, through that platform.

So the Insta cluster developers, they actually participate in the community. One of the top committers in Cassandra is a Insta cluster developer. So, so Insta Cluster itself is not a open source.

No, that's clear. Yeah. Yeah.

So, so yeah. I mean, our plans with Insta cluster is to definitely, you know, we are going to, over the next two years or so, we are going to play a more involved role in the community, uh, today. You know, we, we do in some, we do, don't do in as much as others, but we, we plan to do it.

Yeah. So, so not just like, think of Insta Cluster as a product is a control plane that is running a whole bunch of this open source projects or products as a kind of a data plan. And we also participate in that data plan in some of them, and, and we are going to increase more in others.

Cool. All right. Coming back to cloud storage.

Um, you know, when customers move to cloud, they have options, right? To choose for their block storage needs, file storage needs, object storage needs. There are reasons these enterprise customers, especially for enterprise use cases, are, are, are using ontap uh, across clouds.

You know, when it comes to general purpose file sharing or file sharing for high performance applications, customers really like that multi-protocol nature, right? So you don't have to split, uh, my SMB file shares to my NFS environments. Um, they can, you know, really share the same data across multi protocols.

And now, recently, we've also added, uh, object access to the same file data, truly unifying, um, and, and simplifying, uh, customer's data as estate for traditional file sharing, you know, high performance, uh, uh, machine, um, applications, as well as analytics applications, uh, through S3 without making any incremental copy of the data, while still simplifying, you know, the, the data management, the data security at one single data source customers such as Microsoft, you know, we spoke about, um, in, in our earlier session today, running thousands of cores to run simulations to develop in-house custom silicone. They use the power of a NF customers like arm, you know, utilize FSX for ONTAP to build the next, next generation chip. And these applications require extreme performance, you know, tens of millions of iops, petabytes of data, tens of gigabytes per second throughput.

They also need to burst to different regions or from on-prem to cloud because they want to essentially utilize the compute wherever they can get it, right? It's about the speed of execution. And NetApp really helps them to get the data wherever they need to run their compute, right?

Flex Cache is a huge game changer for such customers where now they can, without worrying about moving large sets of data from one place to another, they can essentially fire up compute instances wherever they are, whichever cloud they are, and essentially start, you know, running their jobs without moving full data sets, right? While still simplifying access to the same data across hybrid multi-cloud environments for VMware environments. A lot of customers are lifting and shifting from on-prem to cloud, but they want to maintain the same consistent VMware experience.

All three hyperscalers now have a VMware offering, and what is common across them besides VMware is ontap, right? We bring the same benefits of running VMware, whether on NFS storage, block storage. We bring the same benefits of data protection migrations, you know, accelerated through our Snap mirror capability.

All these things matter to our customers, especially as they are designing an enterprise grade VMware solution, but now trying to take advantage of the agility, the cost economics, the performance, the scalability of the cloud. Have you Seen a lot of your customers, um, using, uh, your connection to the, the private clouds, to, to, to move their implement VMware implementations to the cloud as they figure out a way to maybe, um, decouple from that platform? Uh, yeah.

We have customers who are, uh, taking the, you know, essentially the route of moving from VMware to a cloud native compute, right? And, and this is where, again, you know, when you run cloud native compute, a typical pattern is we need block storage for the VMs, right? And rather than, uh, utilizing, you know, other different block storages, uh, they can standardize on NetApp, right?

And a very typical common pattern is also it's just not blocked. I also need file storage for those VMs, right? So customers get that benefit of a unified storage solution, you know, on, on, um, on, on NetApp, whether they're on-prem or in the cloud.

But to answer your question, yes, that is a, a pattern we have started to see. Um, and being on ontap, especially on-prem, for example, really helps those customers carry their skim skillset, right? The same management paradigm, the same data product protection paradigm as they were used to on-prem with VMware.

And now it's essentially a lift and shift, right? For the lack of better term, you know, in the cloud, it's simple. It's the same, you know, it's operationally the same, um, but at a better scalability factor, better cost factor, right?

Um, better optimized to take advantage of the cloud infrastructure Database is another big one. Um, you know, going back to block storage, um, a lot of customers, for example, in in, in, in WS are picking, um, um, fsx for ONTAP for running their highest performance database applications, which require low latency, you know, hundreds of thousands of iops, um, very strong data protection. And with block storage on ONTAP customers get that multi-zone availability, um, segregation of performance of IOPS and capacity through throughput.

Customers get that flexibility of really fine tuning, uh, the storage component for mission critical and high performance databases, which is not possible in other storage solutions. Let's dig in a little bit more, uh, into our first party offers. Um, when, when we start, when, when we started partnering with, um, the, the hyperscalers, the intent was never to lift and shift our ONTAP platform to the cloud as many other industry players are doing, Hey, here's a software version of our storage.

You know, you just, you know, run, run into the cloud and, and, and claim that, hey, we are cloud native, um, with each and every hyperscaler. It has been a very intentional design tenet to essentially use the cloud constructs and the cloud technologies and their ecosystem, and build NetApp for that cloud so that customers have that native experience and be able to take the late advantage of the latest technologies, the security mechanisms, the tiering mechanisms, the scalability mechanisms of each cloud. So it's just not NetApp running on the cloud as a lift and shift.

ONTAP has been really architected for each cloud, you know, to bring the best of experience, the cost benefit, the performance benefit, the scalability benefit of each and every cloud. Um, you know, mul, I spoke about multi-protocol, right? In, if you look at A-W-S-F-S-X for ONTAP is the only unified storage service store, uh, storage service in AWS, which provides one single storage service for file, block and object.

No other storage service can do that. If you look at a NF, the only storage service that can provide files, N-F-S-S-M-B, and now S3 access as well. If you look at Google, the only storage service can do file N-F-S-S-M-B, and we recently announced block as well, the whole message of unification for simplification.

We live it every day because it really matters to the customers, right? It's not that the, the, the complexities of designing a a A A cloud architecture are simple. No, it is not.

You know, we are playing that role of bringing simplicity, at least at the data layer. Scalab, scalability and disaggregation. Those are another two big, you know, factors for, for our customers.

When customers use, um, ONTAP to really design their applications, they want that finer grain con control, um, on, on performance, on scalability. They want to ensure that I'm getting the IOPS needed for IOPS hungry applications. I'm getting the right capacity and performance, you know, for, for, let's say capacity oriented applications, right?

We provide that flexibility so that customers can really fine tune the storage as per different applications. Cyber release resilience is baked in, um, with our anti ransomware protection customers can, you know, be, uh, reliant on NetApp to protect their data, whether it's file data or block data to ensure that the data remains secure. It's also worth mentioning that, um, it's not only just the product integration, right?

It's also a lot of go to market integration, you know, with, uh, with the hyperscalers, uh, when customers buy one of our, you know, one piece services they're buying from Amazon or Google or, you know, or, or, or Azure, uh, they're being built by them, uh, built by them. They're being supported, you know, by these hyperscalers. Um, the whole infrastructure is that deployed in the well architected fashion?

You know, for each hyperscaler, most of our customers are global. You know, we are spread across 130 regions across all three hyperscalers. And, and, and, and that really matters to the customers because the majority of enterprise customers are all over the world.

You know, they have been built over decades and, you know, multiple different units and acquisitions. So they need to be across different regions, and NetApp can really help them with that. I'm gonna ask a, an ignorant question.

Yes. Um, NetApp on-prem or equivalent is surely offered, I don't know if it's offered directly, but it's surely offered on a managed and or utility basis with Our Keystone offering. Right.

Okay. Um, uh, in what way are you, uh, coordinating price parity or, or, you know, similar, um, phrase, escapes me, but, um, you, you, you normally try to enable customer choice through parody, global, global price parody, which is difficult, as well as platform parody so that they can choose to move into the cloud of their choice and everyone benefits without, you know, without a cost beyond the migrating or the, or the upfront cost. Is that, is that something that you, uh, provide any, you know, guidance to customers or to partners about how do you, how do you manage that?

Yeah, I mean, our goal is not to kind of provide cost parity. Our provide is, our goal is to provide that flexibility to the customers, right? And meet them wherever they are, whether they're OnPrem, whether they wanna, you know, indulge in, in a, in a CapEx model or an OPEX model or managed service model through Keystone or a fully cloud, uh, you know, solution.

A lot, lot of our customers have already cloud to commit, uh, with, you know, the hyperscaler partners mm-hmm. And R one P offers, you know, they might want to get convert credits or, or other parts of their, you know, Amazon estate, um, to storage. And NetApp can benefit with revenue on that, and then they free up other budget for more NetApp, for example, OnPrem.

That's, that's usually the business goal involved. It's so freedom of freedom of movement where there's a, there's a cost change involved or less cost predictability. I mean, that's an issue for them.

Yeah. And that, that's where, I mean, if, so, our one P offers are transacting directly to hyperscalers versus NetApp, right? Yes, correct.

So we also have the flexibility of CVO, which is essentially, you know, transacted through NetApp. Um, so when a customer wants to deal directly with NetApp, you know, we have a lot of that flexibility of CVO or Keystone or on-prem, right? Combine that with, with their commit, which very typically is the case with the hyperscaler partners.

Our customers get the flexibility, you know, of transacting both with NetApp across the products that we sell, and of course, you know, the same technology that, you know, gets sold through the hyperscalers. And I think that provides them the ultimate flexibility, you know, to design their financial model and their spending in budgets. Uh, you know, you know, based on where their, uh, design centers are, whether they're on-prem or shifting from on-prem to the cloud or fully in the cloud.

Do you find that the people who typically make the decisions or who actually execute on these offers in the cloud, are they your storage administrators or are they your cloud, uh, buyers? Like, is it, is it in the cloud? 'cause a lot of times in enterprise, there's a cloud silo, there's a storage infrastructure silo, and of course, there's also a network silo, um, which is involved in this somehow.

Uh, but which ones are typically most involved in the purchase of storage of the cloud guys? Yeah, we started, uh, you know, with of course the storage administrators, right? The people who have known for NetApp, you for, for ages, but increasingly, you know, huge majority of our customers are actually new customers that are coming.

Um, and figuring out the, the value that NetApp brings to them in the cloud, whether these are developers or platform admins or the cloud admins, right? We are seeing a lot of traction with these new personas, uh, who have always been on the cloud, but never really got an enterprise class storage in the cloud. But now they have an option Actually, uh, to quickly on, um, answer you, you know, um, I, we have got requests to have Keystone also, which is our managed storage as a service to account for cloud consumption.

We haven't, you know, we, we are kind of starting to receive so that, you know, a customer has a single pane of glass to getting, if they're going to spend X million with NetApp, like how, how that sliding scale moves across on-prem to cloud. We've got some requests there. So, Yeah, I mean, we, it's easy to, to, to go down the rabbit hole on this because, um, you'll also have sales teams and I, I presume there's an NetApp cloud sales team, right?

Yes. And it's probably matrixed and how they work with those who support managed services, though the sales folks who are going direct, especially with the big accounts and everything, it can get quite, um, quite interesting from a go to market standpoint. Yeah.

And complex. Yeah, well and complex, but, um, the, the, the, I think, um, uh, is it OVS, is that right? Um, sorry, they CBO Hmm.

C-D-B-O-C-V-O. Mm-hmm. I have the easiest name in the world.

I never remember anybody's name. Um, uh, or, or, or words like that was a component that I had sort of missed. You're providing a little bit more of a continuum here, unless there's a little lift and shift.

Mm-hmm. Um, there's the financial element, there's the con, there's the consumption and use and the architectural elements too. They all kinda have to work together, so Absolutely.

And CV is available through all the hyperscaler marketplaces, Right? Right, exactly. So, you know, they can still burn out their cloud, commits through even CVO Or bring their own All, bring their own licenses as well.

So we provide that flexibility. Um, one thing before I, uh, jump onto the other part of the, um, portfolio is all the sovereign cloud, right? You know, with what's happening in the world, you know, with geopolitical situations, the, the compliance, uh, uh, regulations that are coming in, it's very important for our customers, especially in regulated industries, you know, sovereign customers to ensure that they have full control of the data, the data resides within their boundaries.

And NetApp with partnerships with, you know, all three hyperscalers, we are, we are fully ready. This is, this is a very intentional effort, you know, with, with our partnership to make sure that we are able to serve these customers wherever they are, whether it's, you know, through mechanisms like, um, you know, Azure Secret regions or GovCloud and Secret Cloud with AWS, uh, we have an option with both our one P offers as well as our CV offer to really serve those customers who want full control of their data, you know, running in, in the most secure clouds on, on the planet Insta cluster, right? That's, um, uh, that, that's a very important pillar, uh, not just for the data platform, but also, you know, moving on in the, in the AI platforms, right?

Um, and as I said, Insta Cluster provides our customers a fully managed zero ops platform to stream data, um, uh, using Kafka, a hundred percent compliant Kafka managed service, um, um, run their transactional databases with a fully managed Postgres, um, offering, um, build, um, secure and, and, and rich AI pipelines, as well as, you know, a search infrastructure log search infrastructure of the fully managed open search, um, run realtime analytics with, um, again, open source, click house, uh, Conder database and query engine, and finally, also orchestrate, um, complex and long running and fault order and pipelines using Cadence, um, and do that without any vendor lock-in across hybrid multi-cloud environments, um, to really architect and simplify not only the storage layer, but also the data platforms that application developers really rely on to build, you know, uh, the modern applications that, that, that are fueling, uh, you know, customers digital transformation. Just on that, um, it is, is it a requirement that the data all be stored on NetApp or no, it's just no. Doesn't matter where you have it.

No. Okay, so it's independent, not, no. Um, cool.

Let's, let's get into some of the exciting, uh, new announcements we've made. Um, at this event. Uh, I spoke about, um, simplifying, you know, through unification.

Um, and, you know, we, we started off as a unified storage platform. George talked about it yesterday. We're bringing all that capability into, into cloud two.

You know, especially with block storage, we've already had block storage with CVO, we had with the FSX ontap, and now we are really, really pleased to kinda bring block storage to Google Cloud NetApp volumes as well. So customers can now deploy up to hundred eight, a hundred twenty eight terabyte volumes, sub millisecond latency, 160 K iops, all with the same control plane, the same flexibility, um, flexible pricing model, you know, within the Google Cloud platform. We think it's gonna be a game changer for our customers who can now, you know, also bring in their high performance databases, you know, Kubernetes applications for persistent volumes, um, you know, along with their file storage, um, uh, usage for the workloads, all, you know, through a unified single GC NME service.

So, is it, is, is that NVME over IP That's icek. It's all sc yeah. Icek very standard platform that customers know already, right?

Um, one of the key challenges, you know, customers get into when starting with the cloud is, how do I get data in to the cloud, right? And there are many different migration options available to the customers, and trust me, it's not simple, right? When you're migrating, um, hundreds of terabytes or petabytes of data into cloud, it's not always simple.

It can take months and months, not just planning, but execution and ensuring that the migration happened, you know, the way the customers intended uhs. You wanna talk a little bit about, you know, what we're doing with, uh, the migration simplification? Yeah.

So one of the big challenges, uh, with cloud is like you can make a decision to move to the cloud, but, uh, moving your application workloads and your data to the cloud, that's a whole different ballgame. And, and this happens because, you know, oftentimes, you know, your source storage, which is on-prem in this case, say, uh, has probably been there or evolved over a period of time, sometimes decades. And that has, uh, you know, caused, you know, and it has gone through, maybe you have gone through acquisitions, mergers, you know, bought companies that have brought in their own own storage, disparate storage, disparate workloads that have had clear behavior or whatever you wanna call it.

And, and so it's, it often turns out to be very challenging, like, making the decision is easy, but, uh, you know, how do you move is, uh, you know, is not that easy. Now, if, if it is, if you're going from ONTAP on-prem to to cloud, that's a easier, uh, there's a easier way for us to accomplish that, which is using snap meter that you probably know. But if you're coming from a non NetApp or a non ONTAP source, you know, that is quite challenging.

And in order to, you know, and, and there are tools available in the market, you know, it's not that, you know, there are migration apps available, there's are open source, but, but migration apps, you know, from commercial vendors can be very expensive. Uh, and open source, uh, you know, uh, utilities that you have, they don't have the support and the enterprise exit feature, so it becomes a pretty big friction, like when you are coming from non NetApp. So in order to kind of, you know, uh, address those, um, uh, challenges, we are previewing, uh, an application, um, you know, uh, we basically, you know, we are unveiling it at Insight called NetApp Data Migrator.

So the whole point of that is for non ONTAP sources, you should be able to move data or migrate data to our cloud destinations easily and quickly. And we provide a full service migration app, which includes discovery and act, you know, assessment, you know, obviously the movement of data, um, cut over rollback, followed by, you know, chain of custody or proof of identity recording. So that's something that, uh, you know, we, we've been working on for a little while, and now we are unveiling in preview, we'll, we'll, we'll do some, some pilot projects, et cetera, with customers, et cetera, and then, you know, go to GA eventually in a few months.

So that's kind of the plan. So, uh, that's, uh, a key thing that, uh, our customers have told us over many months, uh, and years that this, this is something that was needed. Is it possible to go from on-prem to on-prem, a third party on-prem to NetApp on-prem with this, or no?

I Mean, so right now care about that. Now we have, we have, we have scoped the support to cloud destination, so we have tested with A NF and GCNV, uh, and we'll soon do it for the other two, but there's nothing in the code, uh, that makes it, uh, like it, it's cannot move the data to on-prem, but we are right now, you know, testing and supporting the cloud destinations. Right.

Enough. Thanks. We Follow the Customers.

Yeah. Thanks. Is this a evolution of XCP or Separate?

No, it's, it's brand new, like from scratch. Nice. Yeah.

Talk about simplification, right? Right. At the first step, when, when customers are migrating to the cloud, uh, let's move on, um, and talk about ai, right?

There's no conversation without AI these days, uh, but we are taking a very comprehensive approach, right? From the infrastructure level to how do we really drive AI outcomes, um, at, at the core infrastructure level, uh, we have made a bunch of new enhancements and announcing new, new, new enhancements to make ONTAP as the foundation data foundation for ai, whether it's on-prem or in the cloud. You know, for example, customers want to utilize the models, you know, everywhere.

They want to ensure that they can get the GPUs and the data together across on-prem or in the cloud. And data mobility has been a challenge, and we, we have announced, uh, or bring, we are bringing, you know, our Snap Mirror and Flex Cache to a NF and GCNV, and now the first time to really truly give our customers a hybrid multi-cloud efficient data mobility solution, right? Um, we are simplifying the data management across your file shares and your analytics estate, as well as, uh, data, uh, you know, AI training or inferencing typically, which are done with object by unifying the protocol access to the same data across file and object.

With our S3 compliant API, we've improved the performance continuously to raise the bar on the performance. You know, on a NF for example, we are announcing from 12 gigabytes per second to 50 gigabytes per second throughput, uh, on, on each large volume, thereby helping customers accelerate model training and via tuning. And lastly, we are making the data platform a lot more secure, uh, by introducing features like anti ransomware protection, as well as file access logs, uh, audit logging on a NF, um, and, and also on, on the GCNV side supporting gsn GCNV with Google Cloud's, Azure assured workloads, thereby really, you know, providing customers a very, very strong enterprise class AI data foundation across hybrid multi-cloud environments.

Um, once you have a foundation, you have the data, a large number of our customers like, Hey, we wanna do more with this data, right? We want to use hyperscaler native services, we want to use SageMaker to build the models or Azure AI Foundry to build and train and fine tune the models. But most of the time, leveraging these hyperscaler services and innovation, you know, which is at the, at the cutting edge right in the market, has been challenging because this required customers to move the data from file systems or on-prem to, you know, object storage in the cloud, which creates more copies, you know, another silo to manage another silo to protect.

In the end, it's additional cost and complexity, and we're cutting all that, right? We are now making in place data available to all the latest greatest AI innovation through Hyperscaler AI services. Um, and you know, on the AWS side, our customers are already able to utilize Bedrock q Kendra, EKS parallel cluster to use the data natively to build AI applications.

And we are very, very proud to kind of extend, uh, the native integration in Azure with a NF, um, integration with ai, a range of ai, um, and analytics services in Azure using our object API access. And similarly, in the Google Cloud, we announced, uh, integration of our, um, GCNV through NFS with Gemini Enterprise, you know, which was recently announced by Google. So for example, if customers are building, you know, semantic search applications using AI search or analyzing large volumes of data using Azure Databricks, or utilizing, uh, you know, Azure AI Studio to build fine tune run models, um, or using a KS, for example, to build their, you know, um, AI infrastructure, they can do that all with the flexibility of NFS storage or now also object storage through the fabric connector, um, of, um, um, of a NF and, and Microsoft, uh, without really making any data copies, without creating another data silo, without creating another data point to secure and manage, thereby simpl, you know, simplifying, uh, managing storage while think being able to take advantage of the latest AI innovation.

So On the last couple of slides, you've, you've mentioned various Kubernetes services on there, any thoughts of integrating with Q and multi Q so that I can move data when jobs are scheduled? Um, um, I think that can be orchestrated, right? Um, so yeah, but It needs something, so if it was built into the data ops toolkit or something else, like, so if a job gets scheduled here, yeah, I want the data to be there when the job's gonna run, right?

Right. And, and so some form of, it's something that's come up a number of times in other projects I've been working on where it's like, if I'm scheduling across lots of different clusters using qr, multi q, um, I need to be able to, I don't deterministically know where it's gonna get scheduled. So when it's about to be scheduled, the data needs to get to where it's going as well.

This is where, you know, like the programmatic access to create flex ces, for example, can come really handy, right? Yeah. That would Be a good integration.

Yep. Yep. Absolutely.

That's, that's, that's a good, good thought. I think we'll, we'll take it back to the, you know, product team. Cool.

Cool. Um, similarly, you know, in, in, in Azure, you know, I spoke about Google Cloud NetApp volume integration with, you know, Google Gini Enterprise, right? So customers can really build agents with their enterprise data, right?

Grounded in the knowledge that comes from enterprise data without really making any data silos. Um, lastly on the Insta cluster side, um, as I said, we already have a strong foundation of Insta cluster, you know, to, um, to run the key AI data infrastructure founding blocks like vector databases, like AI pipelines, um, and, and analyzing vast amounts of data. And we are making it better, right?

We are, we are announcing bringing PG vector, uh, which is part of which is an extension of Postgres for customers to really unify their transactional and vector data estate. Now with built on ontap, so you can bring all the ONTAP data management capabilities to a unified transactional and semantic search, uh, AI database, you know, one single managed service. Similarly, we are improving the capabilities to build very rich neural search pipelines with open search, fully managed service from Insta cluster.

And lastly, um, we are really unifying integrating data, um, through multiple applications for the AI agents, uh, and the models by announcing an open source initiative to build, uh, MCP gateway that will bring the data sources from streaming to databases to search, you know, from analytics all through a single unified and secure interface so that customers can, you know, really simplify this com, you know, compounding complexity of integrating data with models. And, you know, to kind of, uh, finish it off, um, from the key takeaways perspective, I want you to, you know, leave you guys with, with a few messages that NetApp is the only data infrastructure vendor in on the planet, you know, that provides a true hybrid multi-cloud data foundation for traditional IT enterprise AI and analytics workloads, uh, you know, across many, many different, you know, um, um, segments, use cases. Um, um, we power the most rich integrations with hyperscaler, AI and analytic services while all doing all that in place.

And we continue to build on this foundation of innovation, right? Our, our partnerships are strong, deep, and broad, and we look forward to innovating with our partners, um, you know, as we go through this journey. So thank you.

I.