What is our primary use case?
We use it for our entire virtualization stack, and one of the big driving factors for us was the support between Cisco and NetApp.
How has it helped my organization?
We don't have to worry about support or the resiliency of the solution. Our previous converged platform was Dell EMC. There were single points of failure that were designed into the system, not the implementation, that we couldn't overcome. The only solution was buying more hardware and scaling it out, which was not the best solution or necessarily affordable.
Also, we save days of work when doing new service deployments. With LUN clones we have a template provisioned for the image of our hypervisor on our NetApp, and we can deploy a brand new hypervisor in under an hour. Everything is scripted. We just clone a template LUN and boot from SAN, so there are no single points of failure. There is no spinning disk left in the data center.
Finally, we have easily seen a 100 percent improvement in application performance over our previous platform. It's been night and day, to the point where one of our two identically-configured data centers was refreshed to the UCS before the other, and we started to see a shift in where teams were deploying things. The workload actually became unbalanced because everyone was favoring the newer hardware - they were noticing that it was that much faster. But that also gave us the buy-in from the executive level to proceed with refreshing the other site.
What is most valuable?
The most valuable features are the CVDs, and the support behind it from both companies.
What needs improvement?
One of the things that I've wanted would be availability of a health status, similar to Active IQ from my converged platform, on an app. I have dashboards so I can see the health of the system when I'm in the office, but when I'm not in the office I can't.
For how long have I used the solution?
One to three years.
What do I think about the stability of the solution?
In terms of resiliency, we've tested the failover by pulling direct cables between Fabric Interconnects, IOM modules, our Switch Fabric, we've rebooted things in the middle of the day and we've never had an outage.
It's very stable. I've only had to engage FlexPod support one time for a driver issue. It was resolved on the same day.
What do I think about the scalability of the solution?
We have had to scale the product. We set up one of our data centers with a single chassis and we've since grown into three chassis, all with no downtime.
How are customer service and technical support?
If you open a ticket, you get a response from NetApp and Cisco on the same page, on the same team. Their support has been great. You actually get a follow-up a day later: "Is everything still good?" That's great.
Which solution did I use previously and why did I switch?
In our past converged platforms, we didn't have vendor support that would work together. That's actually what led us to abandon our Dell EMC solution in favor of FlexPod.
Also, there were engineering oversights with our previous Dell EMC solution. There is a single point of failure in the midplane which we had to replace, to the point where we actually replaced an entire chassis. It required a full outage to replace the chassis. On the other hand, there are multiple midplanes in every UCS 5208 chassis and a scale out into more chassis. And those chassis are a lot cheaper and more affordable than the Dell EMC solution. So there is no single point of failure in the system anymore.
How was the initial setup?
The initial setup was more complex for us because we were using Hyper-V. We had to blend a few CVDs, but with the expertise that they had, there weren't any issues.
What about the implementation team?
We had Professional Services from one of our partners. Our experience with them was great. They had Cisco Certified Engineers to assist with everything.
What was our ROI?
I don't have any data about ROI, but I know we were able to collapse some of our compute workload for virtualization and reduce our licensing count for SQL Server. That saves a lot of money every year, just with denser blades that were available in the UCS platform.
Which other solutions did I evaluate?
We did look at the Dell EMC PowerEdge FX converged platform, VRTX. At the time, we were a major Dell EMC consumer. Since our switch to the UCS, we haven't bought a single Dell EMC product.
We went with FlexPod because the engineering was better, but support was the major factor: Cisco support and NetApp support. And they support the product even after end-of-sale. Dell EMC has a max term they will support a product: for compute it's seven years. So we had a situation where we could buy the exact same, older technology product just to get more support. We would have been buying old tech just to continue being supported.
What other advice do I have?
Take your time. It's no small undertaking to implement a converged platform or to shift to a different one. Typically, when you make the decision on a converged platform, you're making that decision for the next five to seven years. So take your time.
Regarding the Validated Designs, I've set up VersaStacks as well as FlexPods and it's just like a recipe book or a cookbook. You follow the steps and it's pretty difficult to mess it up. The Validated Designs are great. They're a great reference guide to go back to if you're troubleshooting an issue later on as well.
In terms of private, hybrid, and multi-cloud environments, it's great to see because we have a large presence in Azure already. But it's native Azure. There was no tooling to tie it to our data center. Until now. So shifting things to the cloud volumes from Azure Blob Storage inserts a common framework, we can replicate data between the data centers and the cloud. It's great.
As for managing private cloud, we use FlexPod for own internal hosting of our customers' data, so we ourselves operator our own private cloud.
It's also innovative when it comes to compute, storage, and networking. You can use any number of Nexus lines, MDS. I've done setups with MDS 5000s. I've worked on systems from version 1 all the way to current, so I've seen quite a few iterations of it.
I would rate FlexPod at eight out of ten overall. It's definitely a very complex system. We're definitely not making changes in it daily. There is a little bit of a learning curve for a junior admin.
Disclosure: I am a real user, and this review is based on my own experience and opinions.