What is our primary use case?
Everybody calls it three-tier, but we came from traditional Big Iron, so app hosting was on clustered Unix or Linux systems. It was nothing to have SAN, NAS, iSCSI, and all kinds of disparate storage in the environment. When we did a run-off of HCI products, we did VxRail, HyperFlex, and Nutanix. It was only a benefit to realize that there was a storage element that we did not account for.
I was trying to beat out having to have all these products to do VMware. In the long run, there was this benefit where we started to think about what other things we could do now that we had this storage. There were files, objects, or whatever you wanted to turn on, so to me, it was a huge win for no money. It was just available. You just took advantage of it when you needed to.
How has it helped my organization?
We scaled out the second or third cluster in anticipation of removing EMC CLARiiON products at the time, which then became Dell EMC CLARiiON products. We scaled out a ten-node cluster. It had 300 or 400 terabytes in it, so we were able to migrate from traditional SAN-based infrastructure. I got rid of the SAN-based infrastructure. I got rid of the refrigerator full of disks and power cooling. We did not have only storage; we could also run compute on it, and everything runs local. We did some storage-centric nodes for a while. They were fun, but they were just large disk arrays that presented themselves inside the cluster. I figured that it was easier to have FAT nodes with enough storage in them to take advantage of both compute and memory versus having any sort of storage-centric nodes, but you could not run any workloads on them. It was a nice Flex environment where you could move stuff into a cluster because it has a ton of storage, but when you tried to uplift it or were close to it, you discovered that some of these nodes do have limitations. In the long run, it was better to have all the compute nodes with memory storage and an all-in-one compute platform for our use cases. Now that disks are fast and everything like that, it does not hurt in the long run. There is no downside.
In the long run, there was money savings. In simple terms, you had to continue either to pay maintenance for the SAN infrastructure and storage environment and then present that to physical nodes that just get shared out when all of that is rolled into one bottle of wine. Imagine taking two-thirds of your infrastructure away with no impact. It is a no-brainer that you are going to save a ton of money. In the long run, you did not have to convince your customer. It would probably be termed as reinvesting into this product. Do not just buy one and consider it done. Start to do the evolution—no more stack delivery, which is also a benefit. You do not have to forklift anything in. Just buy a few nodes every year. It is easy. It is a no-brainer. It saves tons of money.
When it comes to latency or throughput improvements, most of our infrastructure runs on what you would consider just traditional factors. If we had a database or if people had hosted apps, it all ran in the local environment. We rarely separated the infrastructure where we had a database server living somewhere in the environment, and then we had the front-end web servers living somewhere else. A lot of times, they were just combined, but with the all-flash capabilities that Nutanix has in Unified Storage, if you want to move workloads in or tier it, there are no issues. Nutanix has solved all of the problems. I never experienced any issues. I never heard any complaints. No one ever came to me. We have moved things to it. It is hard to get compliments when you are the cloud for your customer, but very frequently, you get asked if something changed. If you ask them why they're asking, they say. "It is a little faster." When it works a little faster, I pat myself on the back. We have been using it significantly, and we never ran into any problems.
In terms of a reduction in the number of storage nodes due to increased storage capacity per node, by being able to be clustered in the same process, we get some benefits from it as compared to traditional tiering. However, at least in my world, as soon as you make some savings, someone will take advantage of it. There might have been some savings, but I would not know. As soon as there is some space to be had, somebody needs space. The customers like it because any benefits that I somehow garner out of the infrastructure, or as I add, they win. If they give me money and I buy something, they instantly see a benefit. They never argue when I say that it is the time to buy. Nutranix MAKES it easier. It is not me. I just do the implementation.
What is most valuable?
For somebody like me who grew up building what others call three-tier and I call traditional architecture, the valuable thing is that you can now buy it. You can have new deployments and infrastructure up and running with very little infrastructure behind it. If you have networking and you buy the cluster, you are done. The flexibility it provides with growth is a no-brainer. It could be node-based or cluster-based. We have multiple clusters. We DR to different sites. We have now bought specific clusters for high-end compute for data analytics and data lakes. There is nothing that it cannot do, and it does everything better than ITS COMPETITORS.
What needs improvement?
A pet peeve of mine is the integration with the hardware platforms. Each one of them presents its own use case. They have let Cisco on stage and they are bragging about it. Generally, when you buy UCS products, they want you to have FIs. FIs are nothing more than just fancy and smarter switches with a GUI. Now that they are partners with each other, you get LCM. LCM did not exist for UCS before. HPE is similar, but the integration points could be made better.
At the surface level, they have done the handshake, and all those are wins. We are able to pick the platform of our choice and deploy it with the confidence that Nutanix is going to back it, but after having used almost all of the products that they have, except for Lenovo and Dell, I would still go with Supermicro. I would still go with the x nodes all day long. The flexibility is there. We only go to one vendor when something breaks. It just makes life easier. There is no short side to it. It is not deficient in some way based on the other ones. They are all equal, and there is more flexibility at a different price point. I don't integrate with extra partners and have to deal with whatever they have. In the long run, it would be just the devil in the details when it comes to hardware.
For how long have I used the solution?
It was the second cluster we wanted. We started using it in 2017 or 2018.
How are customer service and support?
The customer service from my partner and from Nutanix directly is unrivaled. At Nutanix NEXT 2024, somebody was on stage saying that he treats Nutanix as a part of his business, and that is probably the most apt way of phrasing it. I implement it. I see it as an extension of my knowledge or support, so I get to call people and people show up. I do not call hotlines and put in tickets. It is like having two extra guys on your staff that you do not pay for.
I would rate them a ten out of ten. If possible, I would even rate them an eleven. There is not one bad instance that I can think of.
How would you rate customer service and support?
Which solution did I use previously and why did I switch?
We used other products, but it would have been ad hoc.
How was the initial setup?
It is all on-prem. Its initial deployment was pretty straightforward. We had deployed a few clusters before, so it was not a big deal to get the cluster up, but it was interesting to learn the lessons. You think you have all the storage and you can outsmart the unified storage environment by creating a FAT VM, and then I will give it a bunch of storage space. What you realize is that you do not get the benefits of unified storage without using unified storage. Just because you have storage does not mean that it works the way it is supposed to. You learn those lessons. You think that you can outsmart these guys and realize that you cannot. In the long run, it was easy. Once the cluster was up and running, it was just making sure our licenses and other factors were in place.
It is very intuitive. There is a wizard guide for creating shares, objects, etc. It is pretty easy.
What was our ROI?
The biggest return on investment is that it frees up revenue and the data center space cost of dealing with power, cooling, etc because now I can just put more Nutanix in there when needed. It never outgrows itself. As these nodes get denser and smarter, racks stay the same. When I scale out, I never need ten more racks of equipment. I can just fill out some older nodes, and they are now denser, so I stay in the same footprint. That to me is the biggest benefit.
What other advice do I have?
We are not using Nutanix Data Lens. It is probably something I should do.
We are not using much to protect unstructured data in our organization. Most of the infrastructure data that we get is in huge datasets. We store that on larger storage platforms that are designed specifically to hold petabytes of data and then it is read into Nutanix clusters for analytic purposes. I do not have to deal with it. I am pulling from somebody else who hosts huge data frames.
I would rate Nutanix Unified Storage a ten out of ten.
Which deployment model are you using for this solution?
On-premises
Disclosure: My company does not have a business relationship with this vendor other than being a customer.