NetApp and Google Cloud now offer a first-party managed file storage service, making NetApp the only storage vendor to have its storage operating system offered as a first-party managed service by the major hyperscalers.
Google Cloud NetApp Volumes, released to general availability during this week's Google Cloud Next 2023, provides multiprotocol file shares across Server Message Block and Network File System, support for Windows and Linux workloads, and integration with NetApp OnTap's data management tools on premises and in other clouds.
In this interview, Ronen Schwartz, senior vice president and general manager for Cloud Storage at NetApp, discusses the latest Google and NetApp offering, what future cloud services NetApp is considering, and how the current AI hype may affect the vendor's products.
Editor's note: This interview has been edited for clarity and length.
What's the benefit of a hyperscaler-run managed service versus one run by a vendor such as NetApp?
Ronen Schwartz: [Benefits] come from two reasons. One is related to the business complexity of account transacting and other things. The other one is when customers look for a tighter, lower level of integration to be done between the vendor and hyperscaler.
About a year ago, Google and NetApp basically put our heads together and designed what a first-party product can offer. The storage operating system is still OnTap, so features such as snapshots, cross region replication and so on are similar.
We are fully adopting the two teams together for what the Google product needs to have such as the level of security that internal products will need to the lowest level. All of this has been incorporated deeply into the GA product.
We have expanded the regions that we're covering and integration to some of the newest services, such as AI all the way to VMware, to Kubernetes, to Google Cloud Anthos and so on.
The other part is business processes. We're fully integrated, so a single [pane for] monitoring, support, and the automation of site management is put in the product.
Why are NetApp customers asking for first-party managed storage services?
Schwartz: When it comes to infrastructure, there are two buying patterns.
One is customers that don't know where they are going to use the storage and want the flexibility. The other is customers committing to a cloud. They've committed resources, money and people into moving a set of workloads into the cloud or are creating workloads for the first time in the cloud.
When you're in this second case, infrastructure becomes one of the decisions that you're making while you are implementing the workload.
What we found is that this customer has a strong preference to buy in native services from the hyperscaler. The lower you are in the stack, the closer you are to the infrastructure.
What's been the strategy to make OnTap a first-party offering in the major hyperscalers, and how will it evolve?
Schwartz: From an engineering perspective, what we're looking for is that we are adopting the native, core services in an optimized way.
It's one thing to consume other technologies from the hyperscalers using publicly available APIs or through the basic integration from the marketplace. But to be able to collaborate with the different teams and optimize how we are integrating into the core services is one part of our requirements. I think the second requirement is choosing the core additional services that we want to be tightly integrated into.
Bare minimum, from a business perspective, is full automation or the maximum automation possible for the product both in usage of the product as well as managing billing.
What you'll see from us in this partnership with Google is also aiming at stretching into new areas of co-developed capabilities. Data and the storage that powers data is going to be critical for the next generation workload implementation.
AI is heavily dependent on data. I think you will see us co-investing in these areas and trying to empower customer data both on premises and in the cloud supporting Google AI tools.
Something similar that you'll see us doing with a next generation platform and Kubernetes is how fast and strongly we can empower Kubernetes applications to support stateful use cases.
What Kubernetes support is NetApp considering for that next-gen platform?
Schwartz: When it comes to distributed applications with hundreds or thousands of Kubernetes processes running with their own data, what happens if only 99.7% of your Kubernetes clusters are backed up correctly?
The core data services that you expect from storage will work for storage and for Kubernetes applications together with some basic things such as copy. That way I can copy my Kubernetes environment and the associated data and make sure I'm 100% backed up.
What we see ourselves delivering is not just the storage network with Kubernetes, which has been a long-term investment from us with Astra Trident, but bringing the full management of the environment. What you'll see us also do is trying to bring this full management environment into our first-party offering, so this can be used and integrated to the hyperscaler's Kubernetes offerings.
Collecting and structuring data for generative AI in the enterprise has become a selling point for storage vendors. What are NetApp's current plans for AI implementation or creation?
Schwartz: What you'll see from us is multiple efforts in there. The first one is about bringing data even from outside the cloud, into the cloud to support learning from this data as part of a generative AI mechanism. But you'll see us also focus on simplifying the augmentation of the customer's model and data available [into, for example,] vector databases.
Tim McCarthy is a journalist from the Merrimack Valley of Massachusetts. He covers cloud and data storage news.