Hardware Reusability & Repurposing - Part 1

This is by far one of my favorite features of the overall solution between Cisco and Nutanix. It seems so ... simple. So ... "duh" when you simply read the title of the post and what the feature is, yet, this is a major differentiator for Cisco and one that customers should really pay attention to. I broke this up into a two-part series. Context is important - details even more so. Welcome to Part 1. You can find Part 2 here.

Hardware Reusability & Repurposing

"It's my hardware, I bought it - why wouldn't I be able to use it how I want to?" Simple question, complex answer. Cisco is the only supported server vendor with Nutanix that gives customers a level of flexibility with its hardware without having to bend over backwards while jumping through hoops lit on fire balancing a stack of plates ... and still lose functionality.

When you think about it, it's not something customers have had to historically ask themselves. VMware's (Broadcom) model really spoiled customers for as long as I can remember. If we go way back - we virtualized physical servers (P2V), maybe beefed up the server, installed ESXi (probably ESX back then) on the server, and added it to a bigger cluster. For the most part, everything just kind of ... worked (though I'm sure some of us still have scars). This has been the case with ESXi for decades. Things got a bit more stringent in the early vSAN days. The hardware compatibility list (HCL) was a mile long until VMware realized that there truly was a difference between some storage controllers and different drives, but the general concept with the technology still applied. Take a server, make sure it has qualified components, install ESXi, enable vSAN - done.

Now - there's a cautionary tale here that Nutanix learned: Don't blindly claim that any and all hardware is supported with our software-defined storage. More importantly - if you want to ensure consistent performance and a predictable user experience, limit what you support.

Enter the era of "dedicated appliances" - you can also look up Dell's VxRail as another prime example. These appliances have a strict set of approved hardware, firmware, software, and support contracts. Run anything other than the pre-approved software (and sometimes firmware) on these appliances and you're in a pickle when calling for support.

Why am I explaining all of this? That's exactly what's happening with the NX branded SuperMicro gear that has been the Nutanix de facto solution for years. You're buying a pre-built appliance that can only ever run the Nutanix software. Insert your reason <here> why you'd like to use the gear for different workloads in the datacenter and the answer will be "If you need hardware support, you're not getting it."

Conversely - there's also no support for existing SuperMicro customers to take servers that were purchased through directly through SuperMicro and convert them to supported Nutanix nodes.

The two points above also partially apply to other OEMs that Nutanix supports, however, there are a few potential hoops customers can jump through to get some sort of support but they do come with caveats. We'll get into those in Part 2.

I've read and re-read my above statements and paragraphs many times and it may sound like I'm dunking on Nutanix here. I'm not. Or at least not trying to. It's a conscious decision Nutanix had to make for the product and this is a reality that customers need to be aware of and understand. Unfortunately for most organizations, it just hasn't been a muscle they've had to build over the years to even think about asking this type of question - especially with x86 servers!

Enter Cisco UCS!

What does all of this have to do with Cisco? Cisco is the only server vendor that allows customers to reuse their existing investment in server hardware with Cisco UCS and convert it into fully supported Nutanix nodes. Furthermore, customers can also take pre-built HCI bundles by Cisco and re-use the hardware however they'd like in the datacenter. Example: A customer purchased 50 HCI nodes from Cisco but can't consume them all at once. They can deploy a subset of these servers elsewhere in the datacenter for non-Nutanix related workloads.

The most important point: Both scenarios above (repurpose existing hardware and converting HCI servers to non-Nutanix workloads) is fully supported by both Cisco and Nutanix. Follow the Cisco HCL and you're good to go. No hoops to jump through, no special support statements required, and no functionality loss (automated deployments and LCM processes).

This feature of the Cisco UCS with Nutanix solution provides customers with peace of mind on several fronts:

• Hardware Investment Protection: Safeguard your hardware investment and mitigate risk against potential shifts in data center strategy.

• Business Agility: Deploy a flexible solution designed to adapt to evolving business requirements.

• Modern Operating Model: Facilitate the adoption of a consistent, cloud-like operating model for streamlined IT management.

• Asset Maximization: Leverage existing hardware to initiate modernization efforts and maximize the return on your current infrastructure investments.

Why can Cisco do this? What kind of voodoo magic are they doing? No magic: It goes back to its inception over 15 years ago, the Cisco UCS server platform has differentiated itself with a stateless computing model. This approach abstracts individual server-level settings and enables global management through software-defined pools, profiles, and policies.

These concepts apply uniformly across environments of all sizes, from a small cluster of a few nodes to a large enterprise with hundreds or thousands of nodes. This ensures that all pre-approved and co-engineered settings are consistently applied, standardizing configurations and guaranteeing optimal performance, stability, and security across the entire environment, every time.

Going back to the original question: "It's my hardware, I bought it - why wouldn't I be able to use it how I want to?"

The answer: "Cisco agrees and has done so."

Check out Part 2 for more technical details.