What is our primary use case?
We deliver a lot of different solutions on various platforms, including different HCI solutions and solutions like Nutanix and Cisco HyperFlex and NetApps, later HCI mile. Most of them have been on Nutanix and on Cisco HyperFlex and as well as VxRack. Our primary use case has of Nutanix has been for virtualization consolidation. We are partners and resellers of Nutanix and I'm a consulting solutions architect.
What is most valuable?
Nutanix has several feature sets that we like. For example, everything's core centralized on the UI. You don't have multiple interfaces that you have to jump between like in some other solutions. It's more integrated for the overall management of the infrastructure. The other part too which is very attractive, is the fact they provide an option if you're not leveraging your OEM hypervisor like VMware or HyperV. That was a significant cost saver for us as well as enabling us to look at alternatives to the VMware tax.
What needs improvement?
For now, I can't think of anything that can be improved. They've been pretty innovative and have provided a fairly comprehensive roadmap. I've worked directly with some of the backend TME guys and they're very responsive and have addressed anything that's come up. However, I would like to see better visibility with the main OEM backup integrators to have a full backup recovery from site to site and from site to cloud and cloud to cloud - the full range. The cloud ecosystem for public/private, site to site visibility with a single backup product.
For how long have I used the solution?
I've been using this solution for five years.
What do I think about the stability of the solution?
From a software perspective, it's incredibly stable and portable. The only caution I would give is that since it is a software-defined solution, be careful of the underlying hardware. It's nothing to do with Nutanix, it's a hardware issue. You may have a highly available, reliable software platform, but it's on commodity hardware so you might experience more failures on the hardware because you decided to go for commodity. You need to be careful how you're architecting your solution and your application factors as you build up your data center, and not sell yourself short and get the cheapest hardware on the market to save costs, because that can turn out to be an expensive decision.
What do I think about the scalability of the solution?
I haven't seen any issues with scalability. Most people I know for the most part are very sensitive about the fault domains. So they tend to go off from smaller clusters. You do have the ability to go pretty much as far as is humanly possible, it depends how much risk you want to take, but at the same time the scalability is definitely there. Most systems are capped at where you can scale out cluster-wise of VMware, 64 nodes and HyperV at 64. There are two factors to the scalability equation. There's the storage within each node, which is fine and then obviously the scalability as far as CPU and memory go. You can mix and match your platforms on your favorite vendor, but then you need the ability to go beyond 64 where necessary. We do have a couple of accounts that we've worked with where they have some fairly large clusters and I think that's a great option for people needing that level of scalability.
What other advice do I have?
The biggest thing that I've seen has been the backup and recovery which has been challenging for them over the past couple of years. They partnered with Beam and with Rubrik and Cohesity. They had their OEM go-to's, but didn't initially deliver a very good story for application integration for backup and recovery where they had good copy data management. Most of the OEMs are very good at backing up single clusters for local backup and recovery. That said, whenever true disaster recovery was needed and you're leveraging multiple Nutanix missions site to site, there wasn't the visibility of being able to backup or being able to have the application integration.
I would rate this solution an eight out of 10.