Exploring NetApp Astra and Kubernetes: Innovations and Insights – Infrastructure Matters

Exploring NetApp Astra and Kubernetes: Innovations and Insights - Infrastructure Matters

On this episode of Infrastructure Matters, host Camberley Bates is joined by NetApp‘s Eric Han and Shiva Subramanyam, for an insightful conversation on the new architecture and integration of NetApp’s Astra and its implications in the world of Kubernetes for advanced data management.

Our discussion covers:

  • The origins and development of Astra by NetApp
  • How Astra is being used to advance data management with Kubernetes
  • The architecture of Astra, and its design for Kubernetes-native, including CRDs, least privilege access, and self-contained backups
  • Real-world customer use cases in the Kubernetes space, demonstrating the demand for advanced data management capabilities

Click here for additional details on Astra or start a free trial today

Watch the video below, and be sure to subscribe to our YouTube channel, so you never miss an episode.

Or listen to the audio here:

Or grab the audio on your favorite audio platform below:

Disclosure: The Futurum Group is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this webcast. The author does not hold any equity positions with any company mentioned in this webcast.

Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of The Futurum Group as a whole.


Camberley Bates: Hi, I’m Camberley Bates with The Futurum Group, and welcome to Infrastructure Matters Inside. We are here at KubeCon in Paris, which is wonderful. And the booth we’re here with is NetApp. And you might go, NetApp? Containers? KubeCon? What the heck are you guys doing here? Kind of thing.

So, let me introduce you to our guests here that are joining me. This is Eric Han. He is the VP of the Astra development, which is what we’re going to be talking about today. And I’ve got Shiva Subramanyam, who is also the VP of product development for this area. So, cool. So, first thing I want to jump into is, what is Astra? Why did you develop it? What’s going on here?

Eric Han: Great. Camberley, thanks for having us, of course. And so, Astra simply is that we know that Kubernetes is important. And from a storage and data management point of view, we want to make it simple for customers to consume wherever they are. So, public cloud on-prem, Astra has been making it easy for customers to adopt NetApp storage, solve the data management challenges day zero, day two. And we’ve been doing that for as long as Kubernetes has been around. So, it’s 10 years with Kubernetes.

Camberley Bates: Wow. I didn’t realize it’s been that long. Okay.

Eric Han: Yeah. What Astra has simply done is, we’ve built from our storage provisioning, up to data management. We have a lot of exciting things here, that we’re showcasing today at KubeCon.

Camberley Bates: Yes. Yeah. And I’m really excited to get into those details, because I’ve pre-briefed on all of that, and it was really, really cool stuff that you’re going into. So, one of the things I want to tell the people that are listening in here is that I’ve covered storage for too long, way too many wrinkles in my face. And one of the things I’ve always looked at is whatever new is coming up, how is that screwing up how we manage, store, and protect data? And so, that’s why this is so important is because, just like we went through the VMware world, and it was a crisis for many years, we’re doing the same kind of thing here, where it’s difficult to manage the data, protect the data, et cetera, and it needs some special technology and integration with it.

Eric Han: So maybe I start, and I turn to Shiva about some of the technical innovations, but you’re absolutely right.

Camberley Bates: Sure.

Eric Han: VMware started off with test. And then at some point people realized, “I’m putting persistent work closer.”

Camberley Bates: Yes.

Eric Han: And so, from that perspective, you had backups, you had the ability to do patching, and distribution, and cloning. And that story reappears. Now, one thing I think is if you look six, seven years ago, that Kubernetes was still probably earlier on. You saw it was a DevOps versus IT. But now you see both groups show up together. And from us and the tooling perspective, we need to be able to satisfy both. So, the idea here is, NetApp always has a strong connection with the IT storage industry. How do we make it so that our customers can quickly deploy and have that observability, that visibility? That’s where SPOT, Astra, all these technologies, Cloud Insights, come together.

And when you look at it from a state-ful workload perspective, we want to be able to say, from a DevOps perspective, how do you consume efficiently in NetApp? How do you have that visibility, that default built-in data protection? And that’s where, because of the innovations and the evolution, we’re very lucky that Shiva, maybe Shiva, you could talk a little bit about Astra’s architectural evolution, and things we’ve done.

Camberley Bates: That would be great. Okay, this is also something very interesting. Shiva’s coming at us from a database Oracle background. And I’m like going, okay, so Shiva I hadn’t met before. And so, tell us why Oracle database is so important to what you guys are doing with this Astra?

Shiva Subramanyam: Thanks for having us, Camberley. I started in my career as a database engineer. While I was working on databases, I deployed in different infrastructures, VMware, and other bare metals. When Kubernetes came in, and we wanted to get our databases on the Kubernetes, that’s when I had to go through the learning curve of how to deploy databases on Kubernetes.

Camberley Bates: Okay. So, you felt the pain.

Shiva Subramanyam: I felt the pain, right? And Kubernetes was born out of a state-less applications, and we were one of the first few to deploy databases on Kubernetes.

Eric Han: Was that Salesforce?

Shiva Subramanyam: That was Salesforce. So, I led a platform engineering team, which actually did state-ful application migrations. So, that’s how my transition from a database into the Kubernetes world, and then into platform engineering.

Camberley Bates: Okay. So, tell me about the innovation we’re doing here.

Shiva Subramanyam: Yeah. So, Eric talked about the why we do Astra. I’ll just talk about how we did it, right?

Camberley Bates: Yeah.

Shiva Subramanyam: So, the main three focus points for us, when Eric and I started brainstorming was, when we take the goodies of NetApp data management from traditional IT into Kubernetes, we wanted to present that in a much more consumable manner to the Kubernetes audience, the Kubernetes environment. So, we focused on three main things, Kubernetes ecosystem, security and resilience, and making sure all the personas are being served, while we are doing this. So, I’ll talk about each one of them really quickly.

Camberley Bates: Great.

Shiva Subramanyam: The ecosystem. When we wanted to get into this, we wanted to present the data management into a much more cloud-native manner, a Kubernetes-native manner. So, we used a lot of the concepts of CRDs. We integrated deeply into the ecosystem of RBACs. So, we wanted to provide customers who are migrating into Kubernetes, use all the nuances that comes with Kubernetes, not a long learning curve. At the same time, use Kubernetes much more efficiently. So we wanted to build an experience for our customers, while coming into Kubernetes, and use our NetApp’s data management software in Kubernetes. The second point was, we wanted to build security and resiliency from the ground up. So, we built a lot of least security privilege, a least privilege. We built in RBACs, which are deeply tied into the Kubernetes software, into our CRD models. And we also-

Camberley Bates: So, that RBAC is tied into the Kubernetes side of the house, not just on the data management side, is what you’re saying?

Shiva Subramanyam: Yes.

Camberley Bates: Okay.

Shiva Subramanyam: Our data management kind of integrates deeply within those Kubernetes ecosystems. So to provide those secured software for our customers. And the third thing we wanted to make sure is, how do we present to our developer community DevOps personas? We wanted to be semi-opinionated. We wanted to play on our strengths, which is storage management and data management, but let the customers, or let the developers, choose what toolings they would want to use.

So we want to build our experts within the storage, but at the same time provide the customers an ability to choose the toolings that they would like to want. So, we build this model of architecture, which we termed it Architecture 3.0. It’s the third version of the architecture, to just serve these use cases.

Camberley Bates: So, when I went through the architecture piece yesterday on what was going on, I was really struck with what you guys have done, and taken it and put it right into Kubernetes. Correct? Whereas a CSI is a very different kind of approach to possibly, but you’ve still kind of advanced what you’re doing, above and beyond what the CSIs are doing, to a higher level capability.

Almost to the point that you’re almost like, we have these things called container-native storage offerings that are out there. But this looks like it is designed like it’s almost completely container-native storage device, right smack into the overall Kubernetes space. Am I reading it correctly?

Eric Han: Yeah. I think that’s well said. I think if you look at CSI, CSI is great. It gives people the ability to provision and consume storage. But especially let’s talk about once you get into data. Data in Kubernetes itself, you want to be Kubernetes-native. And that’s what we’re talking about with this evolution. And that Kubernetes-native really means, custom resources just means I can extend. And if I extend it into the Kubernetes control plane, it makes the IAM, the RBAC, the security, native. So now a Kubernetes administrator can come in with kubectl, the command line. They can come in with an API from Kubernetes API service.

Camberley Bates: What they already know and know well, and they don’t want to go any place else.

Eric Han: That’s right. And then because now we’re talking about data of the application, we have the pod, the pod specs, the metadata, the secrets, but we’ve tied that together, and made it native secure, and we tie it to the storage. Storage being CSI means that it’s external. And that external part is always going to be the case, because the life cycle’s different. But once you talk about data, you want to be inside the Kubernetes control plane. And that lifting upwards is what’s unique. It’s different than what the industry’s done in the past. I think everyone’s going to have to think about, how do they adapt to this? Because now, every workload is going to run in Kubernetes.

Camberley Bates: So, you used a phrase called custom resource definition. I want to translate that for the guys that don’t know Kubernetes that are the IT op people. And what I see that as, is a deep integration within the Kubernetes space.

Eric Han: That’s right.

Camberley Bates: And so, the guys that are listening into this, CRD, custom resource definition, is something that you see that the guys here at Kubernetes talk about all the time, about how do you take whatever open source product or whatever the latest thing is, and bring it in there, so you have that capability of using the tools that you know best, whether it’s Terraform or those kinds of things, as opposed to going outside to external tools to utilize the data. And that’s kind of a pretty slick way to integrate it.

Eric Han: I think so. CRDs are historic. They’re very powerful. And it’s kind of like an extension mechanism. It’s saying that if Kubernetes understands containers and pods, what happens if I want to introduce a new noun? And in this new noun we’re introducing is around data management, or nouns around data management. But CRDs are also the way people add custom behavior. It could be for things like AI. So, the idea really is, custom resources are extended. They use the Kubernetes API’s tooling, but you could define something specific to your use case. And here we’re talking about data management. And that makes us much more adaptable to storage administrators’ needs. But it also makes us integrated to the workflow Kubernetes, because we’re making it specific.

Camberley Bates: So, another piece of this is that this is, I don’t care where you’re at. I can be on-prem, I can be in Azure, I can be in AWS, I can be in GCN, right? And you’ve got that data management because of your first-party relationship as well with all three of those cloud providers, and your operating system. So, that’s really slick. I mean, when I’m thinking about that, because I may spin up a Kubernetes environment, EKS, up in AWS. I may want to spin it up over there. And now I have my data wherever I need my Kubernetes, or vice versa, or however I want to work that.

Eric Han: And I think this is where, Shiva, I’ll turn it over to you. I think this is where…

Camberley Bates: That’s where he’s smiling big time. I really like this.

Eric Han: He’s stealing all my favorite questions. This is where Kubernetes itself lends very well to NetApp’s strategy, because NetApp is a multi-cloud hybrid. And we’re very strong on-prem. We’re very strong in the hyperscalers we invested in a long time ago. And so, what’s been one of the things that are interesting is containers and Kubernetes make it portable. But at the same time, you want the best storage possible.

So, enterprises are saying, there’s credit card processing companies that are saying, “I don’t use NetApp on-prem, but I’m going to go to this hyperscaler, Google Cloud. How can I get your kind of storage, even if I didn’t use it before, but how do I get it where I’m going in cloud? And because of your nature of being first-party, how do I consume it?” And they standardized on Kubernetes, because they’ve done acquisitions, they want to have a platform to commonly integrate with, and it gives them the ability to centralize, in this case, and move to a public cloud. There they need the best storage. And because we are investing in innovating in a public cloud with a first-party, but also with Astra and Kubernetes, it makes it a very simple choice for them.

Camberley Bates: So, one of the things is the data services that you guys bring to the table. So, ONTAP is very rich with y’all’s data services. In particular, what I understand that you’re doing for protection, or for data management, metadata, and the data, on what you’re doing that has to do with regulations. You want to talk about that, and how you’re applying to like GDPR, and what locality, et cetera?

Shiva Subramanyam: So the way we approach the customers who are coming with security requirements, and wanted to make sure they want to have a region-specific data privacy laws, we wanted to provide them a consistent experience for those customers who are moving between cloud providers, they’re moving between on-prem, and they’re moving between the clouds.

So, the way we wanted to approach them, a much more bottoms-up approach, providing them with data primitives. And the primitives are the same across the cloud providers. The tunings could change. So, we wanted to provide that experience, so they can transition their security practices, and don’t do a lot of tech debt doing the cloud providers. So that’s the way we approach it, from a bottoms-up approach. Eric can fill in from a compliance standpoint.

Eric Han: Well, because we’re in Europe right now, we’re in Paris, right?

Camberley Bates: Yeah.

Eric Han: Just a week ago they passed a EU regulation act. And so, what we’re seeing is, containers, whether it’s AI, the EU regulation act is around AI, it’s really that people want to be able to control where their data is. So especially because you’re asking us, and we’re talking about cloud as well, in the cloud, we want to be able to be specific to that region. Also for us, this approach with custom resources, we can do things like self-contained backups. And that self-contained allows us to make sure that all the data, all the metadata, is local. And whether that’s on-prem, or whether that’s the cloud, we’re creating the abilities to make sure that AI running in Kubernetes, Kubernetes itself, the data, we can keep it local to where that region and that policy is. So, those are examples where these things are coming together.

Camberley Bates: Wow. So, what’s next? I think we touched on all the major pieces here. What’s next for where you’re going with Astra? What could we expect to see?

Eric Han: I think we’re just at the start. I know that there’s been a lot that we’ve done together. I think, from a customer point of view, they’re using us in financial services. We’re starting to see AI on top. So, there will be some innovations in terms of how we satisfy labeling workflows for training. On top of that, I think for us, evolving it so that the hyperscaler consumption is very native and seamless. We had AWS with our FSx Group’s announcements last year at re:Invent, where Astra is available for FSx ONTAP. You’ll see more and more of that into the other hyperscalers as well, those kinds of integrations.

And then also increasingly the data management, we want to make sure that these workflows are simple, easy to consume. And that means that those custom resources is just at the beginning. So, we’ll be even more able to do a lot of these integrations. More you want to add?

Shiva Subramanyam: Yeah, no. You hit the right point. So what we wanted to do more and more is, we wanted to be working with the ecosystem more and more, with the Kubernetes ecosystem. And customers who are migrating their applications into cloud and into on-prem, using the Kubernetes as the ecosystem? We would like to provide consistent experiences for them. So what we are trying to achieve from a bottoms-up approach is, focusing on global and local security. So, when customers move, part of the application that their security is carried over across the regions. And also we want to provide an experience of the data management in CRD model that we adopted, is to ensure that they can use those primitives much more efficiently across the solution.

Camberley Bates: I mean, I love the strategy here, because there’s a couple of things. One, meet the developers, the Kubernetes people where they’re at, which is enabling them to use their tools. Meet the IT operations people where they’re at as well. One of the things that we do see a lot of these IT guys going through right now is, am I standing up a separate cluster, or am I going to put it on top of my VM world? Who’s going to manage these pieces? And they’re already stretched really thin.

So, this gives that efficiency across multiple levels of organizations. You already know this, and this is why you developed it and everything else. But for somebody listening in, it’s kind of like, this is really very innovative. And you guys should be very, very proud of what you’ve been bringing out to the market.

Eric Han: Thank you.

Camberley Bates: Thank you very much for joining me, and thank you very much for listening in.

Author Information

Camberley brings over 25 years of executive experience leading sales and marketing teams at Fortune 500 firms. Before joining The Futurum Group, she led the Evaluator Group, an information technology analyst firm as Managing Director.

Her career has spanned all elements of sales and marketing including a 360-degree view of addressing challenges and delivering solutions was achieved from crossing the boundary of sales and channel engagement with large enterprise vendors and her own 100-person IT services firm.

Camberley has provided Global 250 startups with go-to-market strategies, creating a new market category “MAID” as Vice President of Marketing at COPAN and led a worldwide marketing team including channels as a VP at VERITAS. At GE Access, a $2B distribution company, she served as VP of a new division and succeeded in growing the company from $14 to $500 million and built a successful 100-person IT services firm. Camberley began her career at IBM in sales and management.

She holds a Bachelor of Science in International Business from California State University – Long Beach and executive certificates from Wellesley and Wharton School of Business.


Latest Insights:

The Six Five team discusses Marvell Accelerated Infrastructure for the AI Era event.
The Six Five team discusses Google Cloud Next 2024 event.