15 years is a long time in the world of technology. However, FlexPod is still relevant. In fact, 2024 was the best year yet for Cisco and NetApp’s joint offering, we hear from Kristian Kerr, who is responsible for NetApp’s partner organization within EMEA and LATAM. It seems like FlexPod never gets old. We took a closer look at this platform during Cisco Live to learn more about it.
Pretty much anyone who has ever been to a NetApp or Cisco conference will recognize the device in the photo below. FlexPod always has a prominent place on NetApp’s booth, even after 15 years. So hasn’t NetApp developed any other new and interesting products in all that time? Of course it has, but FlexPod is a very concrete full-stack offering from Cisco and NetApp. In addition, its size (a full rack) makes it stand out nicely, making it a nice starting point for conversation with people walking by.
FlexPod not only looks physically interesting, it is also an example of a successful collaboration between two IT vendors. As such, Kerr calls FlexPod the result of a “very important relationship” that Cisco and NetApp have. The concept has been on the market since 2010 and Cisco and NetApp have created more than 240 so-called validated designs (more on this below). Thus, entirely in line with the current trend, there is also a FlexPod AI. If we are not mistaken, that has been around for quite a long time, by the way. We seem to remember a Cisco Live in 2019 or 2020 where we talked about this with someone from NetApp as well.
The reasons for FlexPod’s longevity and continued success are diverse, we hear from Kerr. “There are validated designs for many use cases, there is constant innovation in the platform and there is support from a single source despite being a multi-vendor offering,” he summarizes some of the reasons. However, there are also some somewhat more fundamental reasons. We’ll address those first below.
FlexPod is Converged Infrastructure (CI)
To understand why FlexPod from Cisco and NetApp has endured and prospered for so long, it is important to start with the underlying concept. That is, FlexPod is an example of converged infrastructure (CI). This is a full stack of network, compute and storage delivered as one big finished product, so te speak. The storage component is from NetApp, network and compute from Cisco (Nexus/MDS switches and UCS servers, respectively).
FlexPod is not only a complete stack in a rack. The individual components work together in the best way possible. This is where the aforementioned validated designs come into play. For each use case, Cisco and NetApp partners who deliver FlexPod to customers can build a stack. Hence, there are now more than 240 such designs. NetApp has created validated designs for several variants of its storage line (AFF, ASA) over the years.
CI vs HCI
With the rise of HCI (Hyperconverged Infrastructure), CI may have a somewhat stale image. After all, HCI is the variant in which the integration between the various components is more advanced. It is also a software-defined offering, where CI is fundamentally a hardware offering. With HCI, there is a software layer that allows you to manage everything centrally and in an integrated way. In the case of FlexPod, there is Intersight for the Cisco components and ONTAP for those from NetApp. It is possible to do some basic storage management in Intersight, for example, but if you want to go a little deeper you will still have to use ONTAP.
However, the fact that HCI is there does not mean that CI has no added value. HCI is undoubtedly easier to use and manage in the basics. However, because of the deep integration within HCI, it is also a more rigid stack. CI is modular, so you can replace or extend components separately. That means you’re more flexible and, in principle, it is easier to scale. With HCI, scaling often means adding an extra node right away, even if all you really want is more storage. With CI, you can basically scale storage separately from the rest.
Kerr also sees the ability to scale up relatively easily as a big advantage of FlexPod for AI purposes. He sees FlexPod as a starting point for what organizations want to do with AI, albeit often after the first steps have been taken in the cloud. That depends on the use case, for which Cisco and NetApp must put together a validated design. If there is one, FlexPod is an excellent place to start. After all, the components are aligned, so the compute and storage can deliver the highest possible performance in the most efficient way possible. After that, it’s easy to scale up. Finally, Kerr points out here the built-in security there is in FlexPod. This is obviously also crucial for AI workloads.
Ultimately, it is important to recognize that HCI should not necessarily be seen as an improved version of CI. They are two converged variants that can coexist just fine. Based on customer requirements, one or the other is then preferable.
Validated designs for specific hardware, as well as workloads
Kerr specifically singles out the ASA line in our conversation. “All ASA platforms have and are getting validated designs. This is constantly going on, which is a good thing because it [FlexPod, ed.] continues to build and grow.” On the day of our conversation, during Cisco Live EMEA in Amsterdam, NetApp had just announced some new models in the ASA line. These models should serve the lower end of the market. Cisco and NetApp will create validated designs for these as well. NetApp also wants customers for FlexPod from these regions of the market. The reason is simple, says Kerr: “There are a lot of the same use cases there.” Then a validated design for FlexPod is a logical next step.
However, validated designs for NetApp’s various storage lines (and Cisco’s network and compute lines, of course), is not all there is to FlexPod. The workloads that must run on it also get the necessary attention. In part, of course, this is still about the underlying hardware the workloads run on. For example, mission-critical workloads such as VMware, Oracle, SAP, and more generally SQL typically run on block storage and thus ASA. Yet there are also validated designs of FlexPod specifically for VMware environments, for example.
VMware has received a lot of attention over the past years, for obvious reasons. Kerr sees potential here for FlexPod. He calls it an “opportunity to innovate in the FlexPod install base.” 80 percent of all NetApp’s ONTAP systems deal with VMware environments, so the company is well positioned to help customers. “We support them whatever they want,” Kerr summarizes. “If they want to move to an alternative hypervisor, we’ll support them. If they want to move to the cloud, we’ll support them with that, too.” NetApp’s deep integration and support with the major hyperscalers plays an important role here.
Most VMware customers Kerr has spoken to, however, stay with VMware. They often have only one question, partly in response to price increases: “How can I optimize?” That’s where NetApp’s years of expertise in these environments comes in. “We see it as an opportunity to modernize infrastructure using FlexPod, for existing and new customers,” Kerr points out. Then, of course, the validated designs help: “Those give the customer the assurance that they can deploy their VMware environments in the best way possible.”
FlexPod has many more years ahead of itself
FlexPod doesn’t really age, mainly thanks to the continuous stream of validated designs. It’s more of a concept than a product, even though the big rack on the NetApp booth does say FlexPod, of course. However, you can’t walk into a Cisco or NetApp building and say “I want a FlexPod”. Well, you can, of course, but then you don’t walk out with a FlexPod. In fact, a partner of those vendors has to build it for you. You buy it through a CAPEX model or as-a-service.
Thanks to FlexPod’s inherent flexibility, this stack from Cisco and NetApp will stay relevant for quite some time to come. As new products come out, so will new validated designs. Or as Kerr put it, “The theme here is constant innovation in a platform.” From NetApp’s perspective, it fits very well into the hybrid story it has been betting on for some time, thanks in part to its excellent links to hyperscalers. From Cisco’s perspective, it fits well with the development of stacks that it has been betting on in other areas as well. Consider the announcement of AI POD a few months ago. We certainly haven’t heard the last of FlexPod.
Also read: NetApp is building an Intelligent Data Infrastructure: what is it?