My name is Nick Howell. I am the global field CTO here at NetApp, specifically for our public cloud services business unit. And that means it's pretty much my job to make sure that everybody in the world knows what we're doing at NetApp. Uh we're not just doing storage anymore. You know, uh news flash. Uh excuse me for just one second. Yeah, Chip, I'm in the middle of presenting. What What's up? What do you mean the website's down? Well, I mean we can move it up to AWS real quick if you want. Yeah. Can give me like 30 minutes and it'llbe up there. >> All right, I gotta go. Bye. Bear with me one second real quick. All right. Uh I got a stupid website sales guys, right? Create stack. >> Who loves YAML? Everybody loves YAML, right? We're all infrastructure people. >> Love YAML. >> XML. We'll go next chipswebsite there. name chip and we don't need any of this other stuff next. All right, fine. So, we'llcome back to this. It'swe'lldeal with Chip when we get done presenting. Um what a long strange trip it's been, right? Uh this is really near and dear to me because at the end of the day uh we've been very invested in being a part of field day because we love this kind of focused uh discussion and interaction with the delegates and everything and it goes all the way back to even before 2017 but I started here specifically because this is where the changeover started happening. Uh and you can even date it back to like 2012 when I was installing ONAP hacking it into EC2 in a weird way. uh 2014 we came out with the data fabric concept. Uh and then starting in 2017 we had we started presenting these concepts to you guys over the course of many years and we rotted through 2018 and 2019 me 2020 uh we started getting into things like tier one applications running in the cloud on some of the services that we're doing and showing how beneficial that is. Last year we launched Astra and that's come a long way in a year. So, at the end of the day, we've got a lot of stuff. And guys, I'm going to warn you in advance. I brought the fire hose today. So, I want to make sure that I get through a lot of this stuff so that you have an idea because the point of me showing you all of those slides just now that we've been on this journey together is because throughout that process, there's this is GA and this or sorry, this is in beta and this is in preview and this isn't quite ready in this region yet and all of these little nuances, right? Everything I'm going to talk to you about today is at least in public preview.All of this stuff we're about to go over for the next hour is at least in public preview. 95% of it is generally available and you can go use it today. So lots of stuff that's been happening over the course of the last decade. Um this the if we just cite the last couple of years especially 2020 we saw Spot uh Talon and Cloud Jumper join NetApp through acquisition. Uh we also last year uh launched uh Amazon FSX for NetApp on tap or I should say AWS launched Amazon FSX for NetApp on uh and we also launched Astra control orthe Astra suite ofproducts I should say that started as one thing but has now become its own little mini portfolio in a way and we'll go over a lot of that stuff today but across the board we're going to go over everything from the foundations of storage all the way up to application management automation through various integration points as you might have just seen me give a quick little demonstration and I'll go back to that after the presentation and we'll show you guys what's been going on. Who would have thought it that a box vendor 10 years ago would have something like this?is thisgives me all the pride in the world, all the feels because this is what we've been hard at work at for five years now. And they it's amazing that we've gotten as far as we have in the last five years. And for the purpose of the demonstration, if you guys want to see all of this stuff actually in practice, this is just a screenshot from our website cloud.netup.com.Each of these things that you see are clickable pages. You can get a full description of each and every one of them. You can get a full demo, free trial, nonothing required. All of it is out there ready to go almost in a consumer model kind of way. This all really started about five years ago when a combination of things started to happen. uh we had the cloud providers coming to us asking us how to do storage better in the cloud. We had uh our own interests of getting into the public clouds and this kind of meeting of the minds came together started out by putting some arrays um in Equinex colllocating things there crossconnecting what we realized was that man [snorts] very quickly was that paz and managed services were going to be the future right if not eventually SAS complete SAS but people are getting away from turning knobs on infrastructure they're ready to get to a point where they can deploy set settings like I did with a YAML file, right? But this be the beauty of what you see on the screen here is that the relationships, the things that we've built over the last five years, uh AWS is sitting here in the room with me while I present today. Like five years ago, that would have been what why is why are they doing that? So therelationships that we've built over those years is really the key takeaway here. But across all three public clouds, we have storage and data management services that were chosen by the big three public clouds. That's important. They chose us, right? Sold, supported, and built by all of those cloud providers. Not NetApp. Not net app. This is stuff that the cloud providers themselves are doing. Right? doing. Right? doing. Right? Speaking of chosen, I can remember a time five years ago when this slide was all zeros. This is again one of those, if I'm beating my chest, I'm sorry. We've been hard at work for the last five years, and it's been three years since I've seen you guys, almost three years. We've been busy uh since the last time you saw us.So, let's go over what the strategy is, how we got to where we are, what's new, all of that kind of stuff. And I'm going to take you down deep down the rabbit hole on the entire portfolio. So, right, this really starts with kind of a mantra of ontap everywhere. What do I mean by that? Well, about 20 years ago, NetApp kind of innovated around bringing unified storage together by putting block storage and file storage on the same arrays. Well, whatif wedid that for virtualization as well with shared storage 10 years ago and then now we're kind of doing the same thing for cloud. You can now deploy a virtual instance or a managed service of ONAP in all three public clouds and have block and file all in the same one. And it can even manage object beyond that. So you now have the unified storage story in the cloud as well. >> Nick, can are there any services that are running in the cloud environments that could come back on premises but don't exist on premises within the on premises offerings? >> That's kind of the point. That's it's I love the way you phrase that. So I'll give you a realuse case. I was at reinvent a couple of months ago back in Thanksgiving time frame and I had a customer come up to me and say I had never heard of NetApp. I had never heard of ONAP until we stumbled across FSX in AWS. We had other threeletter vendors uh in our data centers, right? We had never even really paid attention. >> The minute we saw FSX and how easy it was, we immediately threw everything else out and deployed that app because on tap everywhere, what does that unlock for you as an end user? It unlocks that ability to take advantage of all of our flex and snap words, efficiencies, all of that stuff. So when you build into that kind of single unified platform, the value outweighs thecost savings you might get by going from a >> I'm just thinking that you mentioned some of the technologies that you've acquired and companies you've acquired, right? Maybe some features of those technologies that might actually help people with their on premises deployments especially if you're deploying that technology as a service on premise premises rather than >> buying a box and putting it in. So there may be some things that you've learned from buying thecloud-based solutions that are going to help you deliver better on premises solutions. >> Talon's probably the best example of that, the global file cache. So for edge locations and being able to geollocate data, pre-stage it and things like that on prem, but having a cloud route as a source>> and being able to do that, that's the one that comes to top of mind. >> Um, more cloud, less cost. This really comes down to three things. And as Stephen said at the beginning, um, you may not recognize this NetApp if you haven't heard any of this stuff before. This is at the end of the day we're doing we are doing storage. Don't forget we do still sell boxes. Yes, we do still have them. >> Somereally good ones too. So you can still get our shared storage. And as I was just describing to Chris, what we've found is kind of the inverse of what we have people that are cloud architects for cloudnative operators that are discovering some of our storage services that are above and beyond what the cloud providers are, you know, have themselves natively and then taking advantage of you [clears throat] those services and every the abilities and bringing it all back to onrem with the new on-prem kit. Next thing we're foring into compute [clears throat] with uh spot is probably the best example of that where we can help you take advantage of kind of the black market of reserved instances and spot instances that I'm finding arehard for people to really get their heads around how that works. But spot brings to the table the ability to um manage all of those for you and save as much cost as you possibly can. >> Oh, Nick. >> Yep. >> Yep. >> Yep. help me understand, you know, NetApp wasn't a portfolio company prior to some of the products and when I'm thinking about portfolio companies that are making the transition to cloud, I understand that relationship like I understand that there's a bunch of services that need to be moved to the cloud and I know how to engage those companies. spot Astra on tap there. What's the co coherent message if you're like a traditional cloud architect and you're engaging with NetApp? Like I'mhaving a hard time anchoring how when do I engage NetApp when I have a problem?>> It's a good question. How do you know which product to use in AWS? >> I don't know how many products and services they have in their catalog. So I mean But I know naturally the go I know AWS has like 270 services. >> Sure. >> Sure. >> Sure. >> I know that NetApp I mean that AWS probably has a solution. NetApp has kind of this discovery problem of is this something that I should call my NetApp rep about or is I should I look at one of my portfolio partners? >> That's why we're here today trying to change thatperception. >> Uh you'reright. It is something it the biggest obstacle that we've dealt with over the last few years is that perception, the industry perception. And it's something that I've been on a personal mission for at least the last three years to try to change. Um, and it's one of the big reasons that I'm presenting the material that I'm presenting today because I want to feed you from the fire hose. I want to overwhelm you with the amount of things that we're doing that have absolutely nothing to do with boxes to try and help change that perception so that everybody understands that we're not just doing boxes anymore, right? And it's going to take time. It's a big ship to turn. um and that people have those preconceived notions, right? Uh thank you for that, Keith. Lastly on this uh above and beyond spot and managing your compute and your costs in that way, uh we also had one of our acquisitions in 2020 was Cloud Jumper and they had some really cool technology, some agnostic ways of managing desktops across multiple clouds and on prem and we've now folded that into the spot platform. I'll go over that a little bit later. Called it spot PC and we're also managing virtual desktops. Uh you'd be amazed how much of that relies on performance storage under the covers. Anybody that's done desktops knows how important that is. And the ability to do all of this across a single unified platform is really what comes down and is key. Um everybody has heard of CI/CD before, right? Continuous integration, continuous delivery. Uh there's one more thing I'd like to posit that's really important about this and it's something that comes from the spot innovation side of things and it's continuous optimization. If you're doing CI/CD workflows for your applications, you're probably have different requirements in an ongoing basis as your app evolves and as it grows, as it scales, all of that stuff. You're going to need to take and optimize the underlying infrastructure that your uh stack runs on in order to take advantage the best advantage of that as you can. So, uh I just wanted to throw that out there that it's something to keep in mind. It's not just about when you talk about CI/CD and applications, we're doing our best to bridge the two so that the applications drive the provisioning and configuration of the underlying infrastructure. And that's one of the big things that Spot does. So storage, compute, and even desktops and then some.This really starts with the storage. To me, this is the foundation of the house. And of course, NetApp's going to stand up here and tell you storage rules the world. But at the end of the day, I think all of us haveworked in infrastructure in some way before where we know that you can have the biggest baddest servers with the most demanding apps, but if the storage latency and throughput can't keep up, you're not going to have a good time. So taking a lot of these applications that our customers are wanting to move to the cloud and e either rolling it their own way in a DIY fashion with virtual machines or instances and then or trying to do it through a managed service. They're they were finding that the hurdle was storage performance and cost. So how best can we do that? Well, we came up with the cloud volumes platform uh to be able to give you all of those storage functionalities and throughputs and speeds and latencies that you're used to on prem, but also deliver our rich data management stuff around that at the same time and be able to bring in a lot of that stuff kind of natively or bring it along. We've spoiled customers on prem for a couple of decades now. 69s of availability, uh submillisecond latencies, hundreds of thousands of IOPS. But when they go to the cloud, they kind of bring those expectations along with them. So that was one of the things that not only that, but the ability to have all of their software management, backup, everything still work the same way is an expectation that comes along. So we wanted to build a platform that could do that and we have and it's available across major cloud providers and continues to be available on prem. This breaks down a little bit further into something that I like to refer to as IAS versus PAS. And you can kind of draw a conclusion here. If you're an ONAP customer and you really love all the flex and snap words and you want to take advantage of the things that you're used to on prem, fantastic. We can run on tap in a virtual machine where you can control every single aspect of it and be comfortable. If you've got a dedicated storage team that really likes to get into the weeds of storage and really micromanage everything parts of it, CVO is the answer for you, right? And we can deploy it through cloud manager, which we'll show you guys later. On the flip side of that coin, we've got a lot of app owners and lines of business that really don't care about RAID groups and disk aggregates and fancy networking and all of that stuff. They just want to have high performance storage for their applications. And really all they need is NF an NFS or mount SMB mountpoint, maybe a nice scuzzy one if they are or that fringe. So we've got across all three different cloud providers now. Uh we've have a paz offer that you can just go out and create a volume give it a name how big do you want it to be what service level do you want it delivered at done and you get your NFS and SMB mount point and thatwas the strate we wanted the floor [snorts] to be creating a new volume not have to manage anything else underneath the covers of that. So Nick, uh, so AWS used to have CVS and now I see FSX. So CVS deprecated. >> Yes. >> Yes. >> Yes. >> Okay. >> Okay. >> Okay. >> There are customers on there that we're working with to, um, move them over, but in an official capacity, uh, everything's going to FSX. >> I see, um, Azure, AWS, Google, but, uh, with Citrix Cloud, there's a lot of people who are using Citrix Cloud and it adds a little bit of a an extra layer of complexity. Have you guys qualified this against Citrix Cloud? >> No, not that I'm aware of. >> I was just going to ask the question really around you talk about lands business taking the technology directly. Yeah,>> obviously there's a different delivery model internally when you're a line of business that previously might have taken technology from a central supplier internally gone through some sort of workflow to purchase that and order and all the rest of it. How are the customers that are doing that route now through uh on demand and taking it directly? What are you seeing that they're doing in terms of their internal workflow and actually how are they managing their own process to make sure that there isn't going to be that sprawl and there's still the degree of governance internally within the company.>> It's a great point anda lot of this points to thedelivery model of FSX and the way that they've structured it with uh performance and capacity tiers. So instead of having to pay like per gigabyte prices and all of that stuff and instead of having to get into some of the granularities like that, we want you to have a performance tier price and a capacity tier price. And that'sreally as simple as it gets. The beauty of FSX is that it's kind of the marrying of these two. >> It's kind of the best of both worlds. So it's a kind of a super hybrid of that sits right on that blue line in the middle of where these two land. So you can still get down to FSX admin which is effective and you can get down and control the individual SBMs but you can't get down and like change raid groups and stuff like that. >> So you uh added a FSX for netup on AWS and you removed CVS but actually they were two different products. >> They were >> and so also probably different costs. So how are you going to manage thedifference between the two products because even if you are if you have plans to move some of the customers mean there are massive differences from that point of view. >> So when you say manage the costs uh that assumes a lot that um I have I can't really say in a public setting. Um there's there were reasons that we moved forward with it. there were reasons that we're deploying things the way that we've done with FSX and I'll just leave that as a kind of teaser for the future. Um CVS for AWS was our first foray into doing this kind of thing. It was basically taking storage arrays sticking them in Equinex [snorts] little BGP cross connect and we through the marketplace one-time setup that was kind of a bear and then you could create volumes in your VPC and mount them to your VP uh instances right it's so much easier now not having to deal with hardware that is an expense not having to colllocate that hardware which is an expense not having to deal with networking and egress and direct connects which is an expense all of that stuff matters At the end of the day, that's about as delicately as I can put it without crossing a line. But FSX from a delivery standpoint, from a cost effectiveness, all of that stuff, it just it's win-win win-win-win for both AWS andour customers for that matter.>> Okay. >> Okay. >> Okay. >> You keep focusing on AWS. Is there a reason why you're focusing more on AWS than the other cloud providers? >> Love that you brought that. Let me do something real quick. who was here in 2019 when I presented. Right. I think a good handful of you were. Right. >> You guys might remember this. I brought this guy back. >> Oh, yes. Right. Everybody see that? Hopefully that's in the camera shot. But um Google Cloud, we absolutely love working with Google Cloud because it's at the end of the day it's for me and Imight be stepping out saying this, but it's the easiest one to deploy and the easiest one to manage and is one of the most cost effective ones of the three. >> It's also run by George's brother. >> That has nothing to do with it. [laughter] Um, Azure is still dealing with hardware, but the beauty of it is it's inside of Azure's regional data centers. It's wired up directly to the fabric interconnects. So, you're effectively just getting direct access to a very large storage array in your V-NETs, any of your VMs that you deploy. That's the beauty ofAzure Netup files. So, we've got Google CloudVolume Service that is simply go to the marketplace, click enable, turn it on, start provisioning volumes. In AWS, we've now got FSX where you deployed that. Again, takes about 30 minutes to stand it up. Then you're just deploying volumes inside of SVM's file systems. Basically, you can deploy multiple file systems across different regions. Um, there's one more thing. We got to add one now. So, now, not only do we have the Google Cloud one, but because of the last couple of years of work that we've been doing with AWS, uh, we received this at reinvent last year. I hope we get a Microsoft one soon. It's kind of only up here without them. Um, this is just a testament to the I told you we've been busy for the last few years and this is really whatthis is the pride of where it comes from. Glenda, to your point, why was I'm not intentionally favoring a[clears throat] AWS in any way, but um he is sitting right over here. You guys might not be able to see him, but that doesn't mean anything when it comes to uh where we stand as a platform as for our cloud volumes platform. All of them you can do most of the same stuff with. Um thekey difference to keep in mind is that every cloud all of the cloud providers have their own nuance of how to do certain things, right? In AWS, it's a VPC and an EC2 instance. In Azure, it's a V-Net with a VM. And in Google Cloud, it's VPC with a VM. Understanding all the vocabulary and all that stuff is one thing, but just understand that under the covers, it's the same storage. It's on tap. The takeaway here is that it's on tap regardless of what hardware it's on and what cloud provider it's in. >> So Nick, help clarify some of where you guys are at from a positioning operations perspective. Back in 2014, 2015, it was all about the data fabric. Uh that message has changed a little bit. We're hearing on tap everywhere. >> When I'm listening to some of the cloudnative startups, it's all about operational portability. when I'm listening to the multicloud focused uh storage providers, it's about the same data across multiple clouds. Where's the strategic value in going all in on tap everywhere?>> The strategic value is that you really don't need anything else. The strategic value comes in the things that on tap inherently brings with it. It's not about one vendor's name versus this startup versus their product versus this thing. you all of the flex and snapboards that are native to ONAP work now work everywhere whether you're on prim or in any of the three cloud providers thedoor in the demo after we're done with the presentation I'm going to show you what that looks like >> what value that brings >> I get the op so that to me saying operational consistency across multiple clouds>> what about the kind of the legacy uh NetApp message around the data fabric and having my data accessible across multiple clouds. Is that also part of this strategy? >> It's just an evolution of the same messaging basically. It'smy way of saying data fabric if you will. It's is all it is. >> So and Nick do you plan to uh you know standardize the operational model? I mean FSX is a great product theoretically and doyou have plans to bring something similar to the other clouds because now there is a difference in how you operate thethree services.>> What's the old phrase uh watch that space?>> Can't say but I can't I can neither confirm nor deny.>> Um but to your point FSX awesome segue thank you very much. Um, FSX was launched back at store AWS storage day in August or excuse me, September. Was it September?2nd. September 2nd. That's right. Pretty cool launch. I was really excited to be a part of that one because the things that we were able to finally talk about after two years of building this with AWS. Um, really cool. And as I said, it blurs that line between IAS and PAS. You're getting some you're getting all of the features of a managed service and still getting a little bit of thethings that the storage engineers and infrastructure folks love. the ability to create and um isolate separate SVMs for different workloads and define different performance tiers for different workloads. Some are capacity focused, some are performance focused. Um and that's only going to grow over time. We're going to continue to evolve that as a platform. So, lots of good stuff coming with FSX, but the key takeaway for FSX sold, supported, and built by AWS. This is technically not a NetApp product. That's the big thing to keep in mind. So this is awesome [snorts] that we can now do these things, right? But what stands out to me still is the why. Because a couple of years ago, a cloud purist mighthave said that cloud isn't a place where we should be relying on traditional enterprise data services. So what does your recent experience like increase in uptake in adoption say about what the reality of this situation actually [clears throat] is? What are the barriers to cloud adoption in reality? >> That's a good question. It's about storage, man. Andreally Adam at the end of the day it's you have to understand the difference between the enterprise and what you were describing as cloud arch cloud native operators kind of architect people that are used to working in the cloud versus who would normally come from the enterprise. So I was saying earlier we've spoiled them with a lot of availability a lot of performance a lot of that stuff over the years to be able to do some of that stuff on prem. Well now they want to take everything they're doing and they want to take and just sit it in AWS have it all still work. And unfortunately it's not that simple. So what we've done with FSX is make that as simple as possible. And it really starts with our installed base who are already used to ONAP being able to leverage things like Snap Mirror uh to mirror volumes up to the cloud to FSX seamlessly that keeps all of your storage efficiencies intact. you can start, we have a lot of people that start with backups in DR to be able to stage that almost as if FSX is their secondary data center or tertiary data center uh to be able to replace some legacy gear that they don't really want to refresh anymore. That's where it really starts and it's kind of a bit of a drip feed in the beginning, but once they recognize and realize what they can do, we've seen it just kind of balloon and explode into, oh my god, I want to run SAP here now because the performance isgood enough to do that. All the all of these providers have now acquiesced to the fact that we're now to the point of running these enterprise data services in the cloud. It's now a different world, right? That's >> yeah. >> yeah. >> yeah. >> So the question I have is since this is an AWS service, >> y >> y >> y >> I'm a customer. I go and purchase FSX and I'm running my stuff there and wonderful great and I want to integrate it with my on-rem you know replicate do whatever uh to my on and create a hybrid outcome um support becomes the first obvious question right if I now am having trouble connecting or I'm having a performance problem my replication or my ddup isn't what I expected it to be or you know 10,000 other things would come up I mean you know netapp support is wonderful AWS support is different. Not going to say it's wonderful or not. It's just different. Um, so what happened? What's the experience going to be like for customers that are having issues with their FSX integrated with the onrem? What's that experience going to feel like?>> Yeah, I can't rave enough. Uh, for people that haven't used it yet, the team at AWS that's working on FSX are diehard ONTAP fans. That was a big surprise to me. So, um, Andy and Ed and therest of the engineering team underneath it, this isn't some like sneaky play by AWS. This is like they'reall in on ONAP. Like, they absolutely love it. So, they've built it in as the greater FSX thing that was launched about four years ago >> uh to build out this FSX portfolio. They recognize that ONTAP is number one storage operating system. It's got one some of the largest market share. So, they want to get in the business of enterprise to be able to move some of that stuff up to the cloud. Right. I guess who do you call right?I mean who isnav take service it's whose bill it shows up on >> okay and if it turn you know my only concern is that you know from a point fingerpointing perspective right if the problem you know is it you know it's them it's your on-rem thing and then you call netup support is there any arrangement I guess between netup support and the fsx team to work cooperatively like a >> remember how flex pod used to work bat phone and we would all kind of work together>> so there is an official thing together >> it's not something AWS can solve and theypoint at it and go it's an ontap thing and >> they'll create channel. >> Okay, good. >> Just curious, how safe is it? >> Okay. How safe is it? >> Yeah. >> Yeah. >> Yeah. >> Do you want to vouch for how safe it is? >> I'm happy to talk about it. >> So, it'sreally [snorts] um clarify that because that's kind of a subjective.>> So, not a day goes by you hear, oh my gosh, an AWS S3 bucket unsecured. Something gets, you know, taken from it. >> Okay. >> Okay. >> Okay. So, how does that play with >> from my personal perspective and I'll let AWS speak for AWS. That's usually human configuration error. Okay. >> Making the bucket public when they shouldn't be, >> right? >> right? >> right? >> Agreed. But is there any extra, you know, I mean, >> well, we're not doing S3 stuff, but are you talking S3? Okay. But,that's a good delineation of that. Thank you. >> Yeah. Let me maybe just add to that because thatthat's a good question. I mean if you think about the foundations for any service, this is not just FSX but if you think about what the cloud foundations offers you,have fine grain control with things like identity access management. You have key management s uh or key management uh services that automatically rotate keys. You have cloudatch uh for being able to monitor all of the activity that's going on. And you've got things like cloud trail. All of those have been fundamentally integrated and I guess maybe your answer there is this is why it took two years >> to build from the ground up. >> This is not just ontap software running on an EC2 instance. >> Okay. >> Okay. >> Okay. >> The fundamental difference is this is highly integrated into the entire and wired in to the entire AWS uh you foundational services. Got it. >> So that's why we have that confidence of being able to offer that support. We don't want fingerpointing going on and there is no fingerpointing going on because it'san AWS service that we provide to customers. >> Yeah, we're both kind of all in on this. Thanks for >> clarifying those two different companies, right? companies, right? companies, right? >> So, is it not on tap on an EC2 instance? >> It's on tap but we can't talk about the underlying infrastructure. So, let's talk about how we control these. Now, we've talked about all the different things that we're doing in the clouds. Now, let's bring it all together, right? Because as it was I forget who brought it up earlier there's different nuance in everything of working in each of the different cloud providers and the way that we bring that together is kind of a play on a universal control plane right bringing that together allowing you to control all of your stuff not only your on-rem but also bringing it up to take managing all of your cloud stuff whether you've got it in Azure Google cloud or AWS or all of the above this is where ontap everywhere really becomes real in practice drag and drop functionality full API stack um the ability to control everything you need tomanage the movement and management of your storage across any of the cloud providers, right? Uh really if we break down cloud manager, it comes into these kind of six pillars of things. It's about cloud volumes, which we've talked about uh way too much at this point. Cloud backup, which is something that I'm very excited about, and we're going to continue to expand on that. Being able to take advantage of things like object storage in the cloud to back up on prem arrays, cloud resources, all of that into a centralized indexed repository. That's really that's one of the big things I'm excited about. Second thing I'm really excited about is right next to it called Cloud Data Sense. We used to call this cloud compliance. This is the result of the Cogno acquisition from 201 18 or 19, I don't know, three, four years ago. We acquired a company called Cogno. had a really slick way of just looking at a mount point or a data source and scanning all of the data that was underneath it and being able to identify PII things uh GDPR compliance, CCPA compliance, anything that you might have that an auditor would hate that you have, right? Uh also giving you the ability to look into databases, right? So you now have all of this stuff. This the cloud data sense gives you that discovery classification element. Uh there's a governance arm to it and then there's a compliance arm to it as well to be able to do some reporting for any auditing that you might have to do for regulatory needs for monitoring all of this infrastructure. Werebuilt what we used to call on command insights from the ground up as a SAS application. It's now called cloud insights. Cloud Insights now runs in a browser. uh you have multiple dashboards that you can build across all of the different cloud products that you have uh our stuff and other and it can also monitor all of the stuff that you have on prem our stuff and other. So if you're looking for that central piece of software that can provide you full monitoring customized dashboards you can build an entire knock with nine 65 inch TVs on the wall and have different dashboards for different people to monitor different things. Cloud Insights is extremely powerful and oftentimes slept on. One of the best parts about it is if you're a NetApp customer, you get it for free. Everybody heard that, right? >> Just want to make sure. Now, there is add-ons andpremium kind of upgrades as far as like how much historical data you want to keep and how much you want to, right? But every if you're a Netup customer and you got a support contract with us, all of your you get cloud insights for free to monitor all of your NetApp stuff as a baseline kind of start. If you want to start adding on thirdparty uh managed units as we refer to them, sure there's a little subscription. It's a very consumer style subscription that you can do for like standard basic premium sort of stuff. Cloud taring is a lot of fun. This is one of the extensions of doing what we've done on prem with fabric pools. Imagine if we could now do that in the cloudblock level tiering down to object chunks. That's that gets really fun. And of course, Astro, we're going to go a little bit deeper on uh down the road. But all of this is really driven through cloud manager as a central point. Cloud backup, I kind of just waxed poetic on that. We can skip past that one. Data sense is the one I really want everybody to try. I would love to hear you guys' take on cloud data sense. It's one of my favorite things that not enough people are talking about that we've put out there.>> Yes. >> Yes. >> Yes. >> The ability >> talk more about that. >> Yes. Uh because I'm going to show you guys the reporting. I'm going to show you kind of a dummy data uh PC of what it looks like and all of the reporting and graphs and all that stuff. You can give readonly versions of it to leadership to show and to auditors and things like that to be able to drive conversations around that. But please do me a favor, go try out cloud data sense. It doesn't matter. All it needs is an NFS or an SMB or an object mount point a path. It doesn't need a specific It doesn't need ONAP. It doesn't need a specific kind of hardware. uh cloud insights we just talked about but there was one part that I wanted to make sure of that everybody understood is that the we are up over I think over 550 managed units at this point managed units are the uh entities that we can monitor and build dashboards around so whether that's networking stuff whether that's VMware whether that's your old isolon that you can't stand anymore it doesn't matter plus all of the cloud entities across all three cloud providers all have manage units you can man you can watch your EC2 instances you can watch your V-Netss and your or excuse me, your VMs and your V-Nets and Azure, all of this stuff works, right? This is where it gets really fun. Uh, and I don't know how Kubernetes savvy this audience is, so I'm going to take it. I'm going to assume that everybody because this is Cloud Field Day, right? Weeverybody's heard of Kubernetes at this point, right? Probably too much. One of my favorite parts of Cloud Insights is something that we built into it to be able to look and do deep inspection into a cluster. So we can drill down into each of the worker nodes and we can drill down into each of the pods. We can drill down into each of the things look at all of the networking and almost like a if there's any kind of bullying scenario that might be going on for resources contention that might be happening uh across all of these. We give you beautiful graphs and dashboards. This is what [clears throat] a cluster looks like being monitored and managed. I think there's about 30 nodes in there that it's looking at. You can click on an individual node and look down, inspect the pods that are underneath it, see if there's any networking conditions. All of this again for free if you're a netup customer. The fact that people are still sitting at cubectl trying to figure this stuff out just blows my mind. >> Can I just qualify that a second then? So if I buy my FSX run through Amazon, >> do I still get this as a non-netap customer for free? >> I don't know. >> That's a good question. Holy crap, you stumped me. [laughter] >> Nobody's ever asked me that before. I think so, but I don't want to say absolutely not >> because ultimately youknow you're promoting all of these things as cloud services, but yousaid yourself that you have customer only coming through that route. >> Let me find out. I I'll get back to you before we leave today. >> That opens up another question. >> Um cloud secure one of the things that's important. So it's great that you can monitor everything, but what can you proactively do about preventing stuff from happening? This is one of the hidden gems and we used to have this cloud secure as an individual kind of its own product. We ended up folding it in because it made more sense to do it that way and bring it into the cloud insights portfolio. What does it do exactly? It's going to monitor activity. Well, what does that mean? That sounds kind of generic. We're actually looking at behavior and building ML models around behavior of access to data underneath the covers. So, what does that mean? Well, let me just show you a picture. You've got archive data sitting there. One of the biggest attack vectors for ransomware is stuff people haven't touched for years. It's one of the easiest ways to get in and start encrypting without being known. We've got two ways to do this now. We've got f policy built into the O on core engine and we can see at the core of ONAP when activity is doing things that are different and not normal. In addition to that, if you're monitoring a particular resource, we can uh with cloud insights, cloud secure can detect thatfile hasn't been accessed in about three years. Why is it all of a sudden getting all of this right activity snapshot just in case? That's what cloud secure does. As soon as data starts getting written, you can see there's a little bit of gap right there. But the where that red line isthat is cloud secure triggering an ontap snapshot. So then we can go figure out what's going on. The amount of data that you save by justtaking that just in case snapshot. It cost you 150 megabytes to take that snapshot. Nothing, right? Nothing, right? Nothing, right? Um but theability to roll back to that in a massive ransomware attack like this,is that's what everybody wants to know what ransomware looked like. That's what ransomware looks like. Old data that's been sitting there for years that all of a sudden has a ton of write traffic on it. And we have the ability now, not only in ONAP, but with cloud secure monitoring it with uh modeling built around it. Right? It's real. Um we also keep a log of all of these as potential attacks so that you can go and investigate it. This is the machine learning figuring out what your behaviors are with your data and what gets accessed on what kind of schedule. Uh all of this is built in uh to cloud secure and cloud insights. Okay, enough about storage, enough about data. Let's get into some app stuff because this is where we're going to start talking about um top- down management of applications and underlying infrastructure and really not having to deal with that stuff anymore. We're going to talk a little bit about Kubernetes. Right. So, Spot was an acquisition of ours that came at fall of 2020. And it's been about a year and a half now since we brought them on. So spot came on and at the end of the day what the thing to understand about spot is today when we go out and we provision infrastructure or we use APIs to let um or YAML files to provision infrastructure driven by an application a lot of that is kind of the ondemand resources that are available. There's different ways that you can buy compute in the cloud and this is really fundamental and important to understand. You can buy the ondemand instances which we all know as EC2 or VMs across the different three cloud providers. But underneath the covers, there's these things called reserved instances that you can kind of like bid on and make long-term deals on at cheaper rates. In addition to that, there's something called a spot instance that is you can get one of those maybe for an hour. And there's a lot of nuance to understand and there's some complexity involved with making sure that it's not going to get yanked out from under you. But one of the things that Spot does is automate a lot of that provisioning. We also predict. We monitor what you use. We monitor your traffic. We monitor all of your utilization to make sure that we have resources available foryou to scale up and scale down your application based on demand. And that's the key thing to understand. The application and the demand and the scale are what's driving the provisioning of the underlying compute infrastructure. Wow, we've gotten to that point. It's auton almost autonomous. But in addition to that, we're doing things like uh development or I don't want to say DevOps, all the ops, spin ops, sec ops, and desktop ops. Anybody heard of desktop ops? That's a fun one, right? So the product family for spot, it's its own little portfolio in its own world, right? At the end of the day, it thecore of it is Spot Eco and that gives you Let me just jump to that one. Spot Eco is working with reserved instances and based on the time that you're going to need that reserved instance, it allows you to set that kind of bedrock foundation uh lock you in for some good pricing on core uh reserved instance resources for your long-term applications. These are your applications that are going to be around your core websites, your core um web applications and things like that. This is not really addressing the bursty kind of nature of some of those things. That's where the next thing comes in. This is where you start getting into spot instances to be able to take advantage of demand. So you might have that reserved instance that's got a certain threshold, but bam, it's Black Friday and all of a sudden you've got a million people on your website spot instances and you might only need them for 24 hours, right? And then after that activity comes back down and things settle back down to where your reserved instance levels are at. But the ability to have both of those there as an ongoing ondemand resource management spot just manages all of it for you and you can choose to have something managed or you can choose to not have something managed on a per application stack basis. Right? So you we can import all of your stuff and we can allow you to manage it at the same whether on an individual basis beyond doing this for VMs and infrastructure.This is where we start getting into microservices and ocean. Ospot Ocean is the ability to integrate at the pod level. We insert an operator pod into the across the entire cluster and we're able to do all of the things that we were just doing for reserves instances and spot instances but now for your worker nodes in your Kubernetes clusters all maintained uh with ocean.is just one of the products in so we had eco to do reserve. We have elastic groups to do some of the spot instances, burst demand and things like that. But now we have ocean as well to be able to do that at the much more granular level within the microservices. We're going to take that one step further.We're now going to do this with Apache Spark.Ocean Wave is kind of aspecific version of it for Spark that lets you do this kind of provision on demand provisioning based on demand. Whether it's a batch job running and you're just doing a big job or whether you've got a lot of demand for analysis. Either way, underneath the covers it's Kubernetes, underneath the covers, Ocean is sticking the operator into your cluster to be able to monitor all that stuff and learn about where your demand is, learn about your resource utilization of all of your nodes, be able to take advantage of all that stuff. >> So, are those optimizations within those execution groups? So for instance, it's the optimization just for Kubernetes, just for VMs instances, but not cross um technology. So not looking at it and going, you're running this in the wrong place. This shouldn't be running as a um EC2 instance. This should be running as a set of containers. It's going to be more efficient. Are you doing that level of cross, you know, cross cutting again? >> I'm not 100% sure about that specifically. Like we're not going hey dummy don't run an EC2 this refactor your app and run it like almost all applications at this point can be broken down and refactored into microservices but we're not doing that kind of cross communication what we are doing is providing asingle inventory tree if you will across eco uhlast groups and ocean to be able to see all of you whether it's one or the other. >> Okay. Um there you could certainly do some evaluation and comparison about would this run better as an installable on a VM or would this run better broken down as a micros service in a pod. There's pros and cons for that but no I don't know if we're doing any kind of recommendation like that. >> Yeah. If you're running spot instances and things are very shortlived they're being built and dropped and rebuilt. >> Yeah. >> Yeah. >> Yeah. >> There might be an argument that there's a better way of factoring. >> Yeah. >> Yeah. >> Yeah. >> You should refactor. >> Yeah. So even with spot, it's folks love spot instances. They're great value, but spot instances obviously can fail in the middle of ajob if uh >> so there's feedback and one of the questions I got on Twitter was how's this different between like agraphana or Prometheus or some other uh scheduler?>> Oh, those are just monitoring platforms. >> So the >> Oh, this we're provisioning infrastructure. So you're provisioning infrastructure which is great. So if I'm not Kubernetes centric and I want to use spot because spot kind of predates a little bit ofa little bit at least they're concurrent. If I wanted to just create a Python script to run uh some batch files to do some number crunching and a spot instance fails, I don't need to be do I the question is do I need to be Kubernetes savvy in order to get back though theXML that says a job has failed because X Y and Z restart the job. Can I create my own uh orchestrator?>> I see where you're going. So there's two answers to that question. Um what one of the things that um net spa by netup brings to the table is the unnecess you don't need to know kubernetes just let it give us your application give us your stack give us your configuration whether it'sa it's a python or a jupyter notebook and wave or spark or any of that kind of stuff we'll pro we're going to take over provisioning the infrastructure that your application needs when it's managed by spot we're going to provision reserved instances and spotinstance and as needed to do that. You're never going to touch infrastructure. We just need your container or whatever >> whatever the job is. Right. >> Right. So whether it's a Python script or a full application stack or apod YAML file that's going to deploy a pod replica set like whatever that is spot can do. >> Yeah. When you do that provisioning does it does spot figure out which region to go in? So, itmight say actually US East is too expensive. Let's put this in Ohio or US West or something.>> I'm not 100% sure, but when you put your AWS account in, for example, uh that's it's going to be tied to that. So, it's going to know what your subscription is based on where your location is. Uh there's other IM roles and things like that under the covers that you've got to configure that kind of dictate and set the path on where it's going to provision that stuff. Uh, a lot of it has to do with where the reserved instances and spot instances were acquired from. So it can we look at we are I'm going to do air quotes and say shopping we are constantly shopping for deals on reserved instances and spot instances for our customers to be able to take advantage of this stuff. So we yes to your point we do keep an inventory of what's available out there and if the latency is not too bad there's comparisons across all of that stuff where we can potentially put it in a different region if it behooves you to do so. We're not going to I don't believe we're going to ever automatically move something to somewhere that might be detrimental to the application running. >> That's part of what I was wondering because that gets to be quite a complicated thing like yeah it I mean I might have instances that are running currently in Sydney. Um >> and yeah, the latency is going to be quite different if it suddenly gets moved to what is cheaper but it you know US maybe I'm not too keen on US East because um you know if anything's going to go down it's going to be US East. >> [laughter] >> [laughter] >> [laughter] >> hate it. There's a new US East in Ohio that's not DC. Um, the other thing is you have the capability to delineate and set like what regions you want to operate in. So, if you don't want that sort of widespread thing, uh, you can go in and say, I only want this stack, this app to run in this region. So, you can do that. Um, we I could give you a whole hour on some of the stuff that Ocean does for Kubernetes and that Wave does for Spark, but I want you to One of our last acquisitions that we made last year was data mechanics. This brings in a whole new autotuning. What if we could look at your Spark jobs? What if we could go in and inspect your Jupyter notebooks, just kind of tweak it and tune it where we found some things that weren't exactly configured or coded properly and make that run better. That's where data mechanics comes in. and the integration with Ocean and Wave to be able to just really drive that for you. Again, to Keith's point, you never have to touch the infrastructure and stuff like that. It's gonna it's going to do it. It's going to make recommendations or you can just go fully managed and just let it take the wheel hands off, right? A lot of times this reminds me of the early days of DRS VMware. Spot PC is the last thing I wanted to cover here really quick. Uh, this is what if we could do all of that stuff that we just talked about, reserved instances, spot instances, but for desktops, huh? Power users that need a lot of processor, a lot of memory, things like that, or data entry users that really just need a thin client in the and just some kind of green screen or something like that. The ability to scale these based on demand and all of that stuff, right? Same concepts. Uh, and the last one I'll tease here is a new one which it really falls into the sec ops side of things. We're we want to do inspection into security. So, and manage a lot of that toautomatically close up vulnerabilities to automatically take advantage andlead you down some of this stuff. Uh, we won't talk about that one yet. I got to get going. Um, this is the where I want to close out today. Kubernetes does bring DevOps to the forefront. A lot of the stuff that we've used Kubernetes for in the past has been a lot of throwaway test dev work. Now it's made it into the mainstream over the last few years. And even all of the service providers now offering full-blown supported managed services, whether it's EKS, AKS, GKE, and we have a lot of people that are rolling their own. They're really taking advantage of being able to do some of that stuff that Kubernetes can come to the table. There's complexity involved with it, right? Look at some of these stats. And this is from both CNCF, their surveys that they do every year. Um and we uh 451 research this is not our data this is coming from external 95% of new applications use containers. 55% of companies report some or full adoption of containers. 78% of companies this is the important one because this is where the differentiator is 78% of companies run or plan to stateful applications in production.This is the one that blew my mind from the CNCF survey a year ago. Do you run stateful applications in containers?55% yes in production stateful. So we can talk about serverless all day long, but when we talk about enterprise, they need persistence. They need that constant data. They need the backup. They need the recovery. They need the access. Right? So you might be switching from an installable on Windows or Linux to a container as a runtime, but the data still needs to persist at the end of the day. This is where Astra comes into play. And I want to preface this by saying we've been doing multi-ter application backup and recovery and cloning for all of the biggest tier one applications for a couple of decades now. In 2008, I was deploying Snap Manager for Oracle and Exchange on my NetApp systems. Right? This is very similar to that. It's just in the Kubernetes world and I'm not going to call it snap manager for Kubernetes or any of that silly stuff, but I want you all to understand that that's we've been doing multi-ter backup and application quesing kind of stuff for a long time. That's what Astra is at the end of the day. We're taking and inspecting underneath the covers and we're managing that application, building protection policies around it in the Kubernetes world. And then once you have a backup of that stack, you can then clone it out with snapshots the same way on has done for decades. There's not a lot of stuff that's foreign here for people familiar with ONAP. It's just got a fun cool hipster name and it's doing a lot of the and it's dealing with Kubernetes. So people often get confused and wrapped around the axle about what it's actually doing. It's [snorts] a multi-ter application backup and cloning tool. For now, this is where things get interesting. You can also do all of this programmatically with REST. Um, and with our web interface and you can also use this new thing which we recently announced. I want to put my sunglasses on and be Morpheus for a minute. What if I told you that you could run an ontap storage array as a pod inside of a Kubernetes cluster? What if I told you that your applications and your pods could have their own dedicated storage array? That's what that to me that's what Astro data store is. So we're now so on everywhere, right? Even in your pods. So now we have the ability to take uh ONAP and kind of insert it into acluster stack. So an individual application can have an isolated storage array all to itself. Regardless of howit scales, regardless of how many replicas it has, regardless of how many things, how many tiers it has across different containers, you can all have individual NFS and SMB mount points going out to just those pods that will never kind of transcend beyond the pod, the stack. That's Astra data store. And here's kind of an architecture diagram of what it looks like. It's in currently in private preview. I'd love to sign some of you guys up that are super Kate savvy and want to play with this stuff. Um Glenn, I got you. >> Yeah. Does this support more of the converge model where the storage layers distributed across all Kubernetes nodes in addition to dedicated storage nodes? >> Um think of it more of where is the so the way to describe it is you can have a cluster with multiple node stacks in it worker pools. So you can have an instance of Azure uh a Azure Astro data store that can work inside of a worker pool and it really depends on what your configuration is uh as to what it can see and what it can. So theability that you have to in this model that's on the screen right here you have a basically one cluster uh that is that it can all of the pods that are in that cluster can share it and be able to move around between it but you can isolate it further if you so desire. Yeah.>> Can I is there data mobility from uh volumes or whatever you call them in Astra data store with external quote unquote real on top environments? >> I'm not sure but it's on tap. So why not?>> Well, that's what >> I don't think we've gotten [laughter] there yet. >> Okay, fair enough. >> So your qu technically canI go in and v move >> orjust snap mirror whatever. >> Sure. >> Sure. >> Sure. >> Right. That's cool. I mean that that's huge. Yeah, the ability to snap mirror across Astra data store instances between Kubernetes clusters and different cloud providers and onrem like a lot of that that's where my head starts going like >> to cloud edge. Yeah, it's really the edge play that I'm veryinterested in that. Yeah, >> you mentioned um configuring um Jupyter notebooks so that they work better um like tuning them and thin clients. Um that leads me to think do you guys have your own protocol or are you still depending on other you know providers like the cloud itself or the protocol like HTML or whatever what are youknow whatmakes it perform better? I don't understand. Yeah, it's the ability to optimize your code in your notebooks. Normally your Python stuff that's going to be in your notebook, right, Glenna? So, it'sno different than where when Spot provisions something, you're normally giving it a uh some kind of piece of code, whether that's YAML or some deployment script or anything like that or a Jupyter notebook in the instance of Spark, right? We're going to optimize the instance of Spark. And where the data mechanics tool tooling comes in is having some of that autotuning capability based on the code that's in your Jupyter notebook to be able to inspect and look at all of that and tune the underlying Kubernetes infrastructure appropriately using some of the operators in Ocean. >> I hope I said all that right. >> I'm Yeah, I'm no spot expert either. So, just to be clear, I uh thisis not one of my specific areas. I it's something I'd love to follow up on with you. um and I can get you to the right resource if you want to go deep down that rabbit hole. >> Yeah, I'm just curious on how the data is transum transported and how it is compiled at the endpoint and at the you know source. So okay that's fine. >> So closing out this is the big bow the big ribbon. We got three different things we can really do for you here. It's it the use cases are really important to understand this and the more the takeaway is the more you operate in the cloud the less you talk about infrastructure the more you talk about workloads and at the end of the day what's important there is understanding the needs of the workload not necessarily over overprovisioning infrastructure so that I'm covered for the next three years that's the big thing that you want to understand at the end of the day we want to leverage the cloud to optimize on prem and lower costs right across the board so There may be facilities that operate in the cloud that are cheaper than hosting it and running it yourself on prem, right? Archive data. Who wants to sit on a bunch of pabytes of archive data and run the power and cooling necessary to do that? Uh when I can just mirror that up to the cloud into an S3 bucket or into a FSX instance or and into something else that might run more efficiently and I'm only paying acents per gigabyte kind of metric uh as part of it. Uh the second part of this is operationalizing and optimizing the public cloud. So taking advantage of all of the things that you might use operationally in the cloud above and beyond the storing and managing of data and things like that. So this is where you begin to look for things like universal control planes. This is where you're going to look for things like centralizing and coalesing a lot of stuff where you might have it in disparate entities. Right now, you're going to start bringing a lot of this stuff together and centralizing it. And you're going to start taking advantage of some of those resources that the cloud brings to bear to optimize your environment and be able to carry a lot of that stuff forward. >> Right? >> Right? >> Right? >> The third one is where you've just gone whole hog. You don't care about infrastructure on prem anymore and you're just full cloud native or if you're a new startup and you've got an app or you're building something that's going to run on one of these, you're probably not running on prem infrastructure. You'reall in on one of the cloud providers. So this is where we're looking at cloud ops and developers the ability whether it's SRRES or whether it's application developers that really just want to write code and they want to let that code determine what infrastructure is going to be provisioned under the covers. That's where a lot of this comes in. That's where we're seeing Astra for their in a cloudnative world. What do they do for backup? They need a tool just like infrastructure people needed for Exchange and SQL, right? So now we have Astra to be able to take advantage of that. They can run it themselves in a DIY fashion with Astro Control or we've got a NetApp managed version of that called Astrocontrol Service that you just pay a monthly service on. We'll take care of it for you. We'll help you support it. We'll help you set up all of your jobs, all of that stuff, right? Uh that was the last one. The last part of this is being able to deliver value sooner. And that's the thing I want to make sure that everybody understands is like when you're an app owner, you really just want to get your code out there for everybody to use, right? You hate the bureaucracy of everything. Youdon't want to deal with a lot of it. And as someone who used to be in charge of managing the queue of departments that wanted access to resources on prim. Yeah, we've all been in that boat at some point and they hate that. They don't want to deal with us anymore, right? They can do in 10 minutes for 100 bucks what we used to make them sit in a queue for six months and charge them a million dollars for, right? They've moved on.