As the Converged Infrastructure (CI) and Hyperconverged Infrastructure (HCI) practice lead at AHEAD, I spend a significant portion of my time evaluating and designing solutions. I wrote my last blog post on Nutanix all the way back in 2016. Back then, Acropolis Hypervisor (AHV) was still fairly new having been launched in 2015, and for most of my customers, it was an interesting project that was nowhere near ready for enterprise workloads. Fast forward to 2018—a few months after Nutanix’s yearly NEXT conference—and AHV is a very different beast.
For starters, AHV has most of the features required for most enterprise IT shops to consider the solution. That is, it checks the boxes of security, features and functions, performance, and certainly, management. Since AHV is integrated into Prism Element and Prism Central, one of the key drivers of it is its simplicity. I’ve personally worked with several customers on hypervisor scorecard exercises where they have put vSphere face-to-face with AHV, and unlike the AHV of 2015, the AHV of 2018 is now a strong contender. Let’s dive into some of the characteristics of AHV.
This section is most relevant to those that aren’t familiar with AHV—and want to understand what it can and can’t do. A non-exhaustive feature list is below, using vSphere terminology:
- Management: Native high availability of all management components
- Management: Full-stack visibility (hypervisor down to hardware firmware)
- Management: ADS (Nutanix’s version of DRS)
- Upgrades: Single-click upgrades of the hypervisor
- Upgrades: EVC support for CPUs from different families
- Backups: Change Block Tracking
- Network: Microsegmentation
- Replication: Stretch and Async options
- VM Functions: Cloning, VM snapshots, VMotion, image/template management
One item that isn’t out yet—but that I expect that Nutanix will it release soon—is the lack of storage VMotion. On one side, Nutanix claims that there is rarely a need to have more than a single storage container, which I agree with, but for situations where you have multiple containers, this feature—along with cross-cluster storage migration—is needed.
At this point, the best (and most common) argument I hear against AHV is the lack of a broad ecosystem—that all the backup, monitoring, and logging tools that organizations use to manage their vSphere environment are mostly incompatible with AHV. This argument makes me think of the following discussions:
- Backups: This is arguably one of the biggest hurdles, as it’s difficult to justify replacing a backup solution if it’s working well in an environment—solely because you want to use a new hypervisor. Luckily, AHV revealed APIs for backup integration a year ago, and to date, their list of supported backup vendors is growing to now include most—but not all—enterprise backup solutions. With that said, depending on what you use, you might already be ready to protect AHV with your current backup solution.
- Proliferation of Tools: Should it be concerning that we have dozens of tools that we use to manage our vSphere environment? On the one hand, having only a couple of tools makes it easier to move to AHV, but on the other hand, if I really need dozens of tools to manage the environment, maybe it’s a sign that I should really be strongly looking at AHV—so that I can get rid of the tools and use native capabilities.
- Replacing Your Tools: Regardless of your position about tools, if you have made significant investments in products and training, careful consideration will be needed when considering AHV.
The AHV Edge
Beyond standard and even advanced hypervisor features, AHV stands to have an edge over its HCI competitors because Nutanix has complete control in integrating AHV into the broader Nutanix intellectual property. This is the same advantage that VMware has to integrate its products together (vSphere & VSAN, for example).
One past example relates to the performance benefits that were made available through a change in the way in which AHV interacts with the Nutanix storage subsystem. Marketed as AHV Turbo or “Frodo,” this feature that was released with AOS 5.5 was only possible because of Nutanix’s ability to modify the hypervisor—and it resulted in storage IOPS/bandwidth increases, while at the same time reduced CPU overhead. Since it’s made available through a single-click upgrade—and without adding any new hardware—it’s pretty compelling.
Today, AHV now provides several unique and native features that are not available on vSphere on Nutanix, and this list is likely to keep on growing with redundant VM management plane and microsegmentation being features that are AHV-specific.
As good as this sounds, this situation is a bit of a double-edged sword. The more features and functions that are AHV-specific, the more they could alienate organizations that don’t want to move away from vSphere. Even if Nutanix claims that it’s not their fault—that VMware doesn’t offer an open ecosystem for vSphere (third-party virtual switches, for example)—the end result is that some customers will feel that they are paying for features they can’t use. It will be interesting to see how Nutanix handles this dynamic.
This wraps up my update on AHV, but I’ll be following up with the second part of this blog series focusing on more updates that came out of NEXT earlier in the summer, along with the latest updates from Nutanix’s internal, yearly sales kickoff.