FlexPod Reviews

Most of my customers have separate server, network and storage teams. Therefore, bringing UCS together with NetApp storage, which makes a FlexPod, speaks to my customers' use cases. We are able to carve it up for almost everything, whether it is for HyperVisor's such as VMware ESXi or bare-metal Linux / Windows servers. My customers love the option to carve it up as they like and use their existing storage investments treating the marriage of compute and storage as a "FlexPod", more a platform. Again, it covers off almost every use case my customers have.

The deployment really depends on what the customer is trying to achieve, in-house IT skillsets, budget and other variables will need to be considered. I have customers who are proponents of everything SaaS with a bit of cloud and customers with SaaS, Cloud and on-premises taking a balanced approach to realize an Hybrid "Multi-Cloud" environment. Others that are pairing SaaS with on-premises, and some, only SaaS with remote workers. I also have customers that remain all on-preemies because of the nature of their data. So, my stance is hybrid multi-cloud infrastructure in the end but I look at all the solutions out there, bring the right pieces together trying to achieve, with my customers a hybrid environment. Lots of stuff in-between and many pause at various places on this journey, whether it is to pivot to new technology previously unavailable, re-assess and or build budget for their "the next steps". Hybrid is where to land, where I think all companies are going to land. There will be many variations like a 70/30 or 80/20 split. We just don't know. It all depends on the customer, use case and workloads!

Most of my customers, if not all of them, have a VMware vSphere ecosystem. FlexPod integrates everything in so you can see it all, full visibility into the infrastructure. This is super important for our customers because they want to use fewer tools in the environment. IT admins, in storage and compute, even networking want to be able to use the same toolsets, reducing or consolidating the tools and even vendors. Then, the customer will achieve more streamlined operations and can more quickly see where problems exist within the environment; reducing the mean-time to resolution.

FlexPod, UCS offers a good GUI with easy management. With the management, you can see the inventory of both the storage and compute. There is good integration here and offers a close single pane of glass of management. Most of my customers go to the NetApp GUI, vCenter, and or Cisco UCSM. However, you can see it all under UCSM (Central / Intersight) or VMware vCenter to manage it all. FlexPod provides easy management with a close single pane of glass with good alerting to see the infrastructure as a whole!

VMware vSphere (Hypervisor's) lean heavily on memory. With Cisco UCS, on the compute side, we can get really dense memory hosts to support many virtual machines. With ESXi, we can easily support 50 VMs per host or more. With the FlexPod configuration, we see low latency and fast storage.

The validated designs (CVD's) are important for exploring technology that I haven't touched or seen in-depth. We use the CVD's to get a better understanding of the technology and use it as a roadmap to get customers to that "desired end state".

A lot of my customers don't take advantage of automation but with UCS, software defined templates, policies and pools are heavily used and save time. Generally, you make templates to help with the automation of provisioning of server, network and storage configurations. Cutting a server from a template, creating a server profile, pre-configures compute, network and NetApp storage. This is super important because it reduces the time to deploy a host and or virtual machines.

The traditional UCS Blades do not take much storage internally. You would be challenged to create an HCI (Hype converged Infrastructure) solution on FlexPod / UCS or any other solution that pools internal storage. Now, with UCS X-Series, you can carve off an HCI solution, software defined pooled solution if you want. This was one area of improvement that I wanted to see and can now realize with the refresh of the Cisco UCS infrastructure.

With modern modular infrastructure, RESTful API has been added, there are more integrations, ServiceNow and vCenter along with tighter plug-ins. There is cross-user interface launching, for example with Windows Admin Center. The solutions are using Ansible and Terraform for deploying infrastructure as code. All the improvements that I wanted from the last gen are here or coming. 

With modern workloads and GPU use on the rise, adding GPUs to modern modular infrastructure will have some pros and cons. Typically, you can add one or two GPU's to a blade with no or little trade off. With the UCS X-Series, if you are doing a GPU farm, then you may have to sacrifice compute blades in the front slots to put in a GPU tray / module. A chassis holds eight compute blades, but if you are adding a ton of GPUs, a single GPU tray or more will reduce your blade count by as many GPU trays you add. This is not just a Cisco UCS X-Series problem. It is an industry problem with modular infrastructure and one that I would like to see get solved! I am looking into one such solution, VMware BITFUSION where you can send CUDA requests over the network to a BITFUSION server with the results sent back to the requestor, early stages here and only scratched the surface thus far.

With Cisco UCS X-Series, I would like to see the fabric interconnects built into the chassis instead of being external. With the fabric interconnects, the real footprint of UCS X-Series is 9U, where some of the competing solutions are 7U and have collapsed the network fabric into the chassis. This is another thing that I would like to see from Cisco, though, not really on the NetApp side of the fence, NetApp is solid storage.

Cisco UCS came out in 2009 and I have been on the scene since. FlexPod came next enhancing the UCS offering with a marriage of UCS and NetApp storage.

I have built over 15+ datacenters on UCS / "FlexPod", and every solution has been solid. I have done multi-domains within some of my customers datacenters too. More of the problems are "people-generated" / "user error, understanding" than they are technology-based. The stability with this infrastructure is rock-solid.

Scalability is the best that you can get. You can carve it up any way you want, e.g., add storage or compute. With the new UCS X-Series, some of the shortcomings in the traditional UCS have been resolved. Now, it is even more scalable and granular, if you want it to be, but it doesn't have to be.

The flexibility, operational efficiency, and scalability of FlexPod is really high. You can add compute and storage independently with ease. There are no real concerning limitations other than having 20 chassis in a UCS domain and then you need to start another one, which is okay, you can then connect all the domains you have to Intersight for that single pane of glass of management. Most customers don't get past 10 chassis at a single site.

I support my customers but we have called Cisco TAC who has been very responsive. Because it is a FlexPod design and marked as FlexPod, we don't have to call multiple vendors for support. We are calling one number and generally getting responses back within four hours, which has been very acceptable. I know we have called support a dozen times over the last couple of years, and everything, for the most part has been resolved within a few days.

Even though TAC responses, in my experience have been within four hours acknowledging the problem, there is always room for improvement. We have had a case or two with some customers where troubleshooting took a week or longer and I have been asked to come in to help. I help the customer by bridging the gap, which is fine. That is what CDW does, take care of our customers, bringing all parties together when needed. This has only happened a couple of times that I can remember in probably the last five years.

I would rate the technical support as nine out of ten.

Positive

A lot of my customers use three-tier infrastructure, like HPE, Dell, and Cisco Rack servers in addition to all sorts of different storage (DAS, NAS, SAN). A lot of my customers also have all of the the vendors and equipment in the mix, e.g., where they had storage from EMC, from NetApp, and from Pure with the same on the compute side with generally, all Cisco networking.

With the FlexPod solution, you are standardizing on a platform and streamlining operations. You are utilizing Cisco compute and NetApp storage, reducing the vendors, less sprawl, fewer admin tools and easier management.

The initial deployment is straightforward. Looking at the validated designs (CVD), you can create detailed implementation plans from.

Depending on the customer, a lot of the time, the customer wants the compute environment build out first and then attach the NetApp storage and flag the environment "FlexPod" after the fact. For SAP HANA, another example, I would look for the CVD and TDI (Tailed Datacenter Integration) so the customer and I understand what the requirements will be, can build an estimate out, architect the solution, put an accurate order together and build it according to the CVD.

I find the CVD's very important and helpful as they provide good reference for the technology that is being used or considered. If a customer has it in hand ahead of time, they get a better understanding of what it takes to build the solution out. Whether we follow the CVD through, at the end of the day, that depends on the customer.

I have actually done end-to-end deployments with my customers along with guiding them. I find it very straightforward. After I go through a few sessions with my customers, they are on board, get up to speed very fast and become very knowledgeable.

For ROI, I would have to account for all the time savings and streamlined operations. I haven't done an official ROI yet with the customers. However, when we compare cloud-spend, we are seeing for three years in cloud, we could have purchase FlexPod on-premises for less. Many customers on-premises are achieving the hybrid solution, they are adding cloud to it when and where it makes sense. As a whole, customers are still liking and buying FlexPod as it is today. On-premises spend has increased on the server and storage front with increased cloud spend.

The templates, visibility, and ease of management in my opinion reduce deployment and troubleshooting from a couple of hours or more down to a fraction of that, thinking of an incident, you will also have fewer people on the calls. You are not getting everybody from every aspect of the business, networking, storage, compute, managers, and PMs onto a conference call just to figure out what is going on. You are getting onto a call with specific people knowing what is going on from a single pane of management and alerting perspective.

Other ROI, the FlexPod solution greatly reduces the cabling in the datacenter returning back some savings by way of physical gear purchase and physical IT management time. We are seeing reduced time to troubleshoot, reduced time to inventory the infrastructure, and reduced time to deploy systems with templates. In my experience, the solution is saving easily 10 hours a week in an enterprise environment.

When customers are standardized on this technology it becomes consistent infrastructure and easier to expand when needed. The customers will also have consistent pricing and know what to budget for expansion and or builds in other datacenters. It means that the customer does not have to go out and re-architect for new solutions to fit a workload with new pricing, infrastructure, components, and then slap it all together, again. There has definitely been savings in standardization of this solution by not having to go through the whole process involving more people and adding infrastructure sprawl that needs to be managed differently with more tools, more vendors, managing many service and support contracts.

Licensing can be interesting with the UCS refresh because of Intersight integration (IMM) and SaaS, connected virtual machine or private virtual machine. I have seen some of the different licensing models, but with the new purchasing, all customers are going to have to buy some sort of Intersight licensing with the new UCS X-Series, "FlexPod". I will explore the licensing, setup costs and options with my customers as we discover, assess, architect and get some estimates going!

I have evaluated most modern modular solutions and deep with Dell Kinetic Infrastructure MX7000 and Cisco UCS solutions. I have explored HPE synergy and Lenovo Flex but feel UCS X-Series and Dell MX7000 are the two most modern modular solutions, each with their pros and cons depending on the customers use case, workloads and desired end state.

Again, FlexPod can be thought of as a platform with easier management and granularity to carve it up for almost any use case, a mega block of modern modular infrastructure and would rate this solution an eight out of ten.

We use FlexPod for onsite virtualized and leveraged hardware, so we work with many VMs and generic workloads. There is no specific single-purpose workload. It would be like the VMware solution.

We have three primary production environments and a legacy environment. Right now, the entire workload of the organization is roughly split into thirds across those various generations: Gen one, Gen two, Gen three, and legacy. Legacy can't move off of old hardware, or it's in the process of being moved to modern platforms. 

Gen one was FlexPod, Gen two was a different Dell solution, and Gen three was yet another Dell solution because somebody up the chain loves Dell these days. They're building a fourth generation of primarily containerized stuff based on Red Hat Kubernetes for new workloads or Red Hat OpenShift. The compute workloads in FlexPod are slowly being moved out. There's no plan to replace that or to continue using that as a FlexPod. However, the storage driving the current compute and network shares will be transitioned into the strategic platform for file-sharing platforms across the enterprise. 

It highlights the flexibility that comes with the product set. They're still using VMware everywhere. They're not going to be using the compute environment. They're still using Cisco network switches everywhere and will continue to use NetApp storage everywhere. Interestingly, this new containerized OpenShift environment they have in place today gets its storage back from the NetApp through NetApp automation and integration tools in their cloud set. It highlights the flexibility of what you can do with the entire platform when you need to do something.

In terms of administration, we have five or six people on the storage side that are primarily doing storage. There's a rather large team in charge of server administration, but they're doing server administration across the entire enterprise as a team. There are around 25 or 30 Windows and Linux system administrators. The house's Windows side also manages the necessary VMware work, but they're not dedicated to FlexPod. They're across the entire enterprise, but any of them could come in and do most of the work.

In addition, there are about 10 people on the engineering side. They're the ones who are primarily responsible for designing and driving operational standards. The same people are in charge of infrastructure across the entire enterprise and aren't limited to FlexPod. There's also a good smattering of network folk. 

The pre-validated architectures are attractive. My organization is in the Fintech industry, where standards are critical, so validations and compliance are essential. There are vendor-designed and vendor-validated design starting points. You can build up for scale or if you need more compute from those starting points. If you need more storage, you put out more storage. For performance and things like that, you can put the components you need to meet your specific needs as long as you stay within those validated designs.  

Even if you start with one and diverge from it, it's still a FlexPod, and you get support from the three primary organizations: Cisco, VMware, and NetApp. Even if you don't use a Cisco-validated or NetApp-validated design, you can still qualify as a FlexPod if you're out of those standard deployments. Executives like having those models as a starting point.

FlexPod has saved us some time, but I must add a caveat. The challenge was getting the operations side of the house aligned with what FlexPod was. Even today, the operational side remains heavily siloed. There are teams for networking, storage, Windows, etc. That separation of concerns continues to be the standard model for operational support today. It took a lot to break down that model a small amount because everybody in FlexPod has to know and work together because of the integration. 

You can't just do the network piece because there's so much that the network touches. You can't just do the compute piece because there's so much the compute touches, especially with VMware. You can't just do the storage piece. You have to understand how that relates to virtualization and the compute. It took them a while to get some people in place that could cross these silos to make troubleshooting more effective. Has it become more effective? Yes. But it took a long time to get there for external reasons.

It hasn't impacted our TCO because everybody who came in still runs everything else they used to run. When they decided to do the next major expansion, and somebody wanted to go with Dell, and they went with Dell hardware, all the same people had to learn that as well. So we don't have a TCO advantage on FlexPod versus anything else.

The pre-validated architectures also haven't affected our productivity if you look at application performance. The application performance is what it is. It also hasn't increased our productivity regarding our ability to operate the system because many outside factors affect our operations, and nothing about the FlexPod design helps us overcome them.

The most significant benefit is storage flexibility. FlexPod is a reference design to build on and add the appropriate resources you need for your workloads. That's what differentiates it from something like HCI, where you have to buy certain size pieces when you want to expand.

I can do just about anything storage-related with it. In a FlexPod environment, we can provide whatever you want from a data storage perspective. Replication, backup, disaster recovery, etc., are all there right out of the box. I can set up fiber channel LAN, channel-over-ethernet LAN, etc. Storage is highly flexible in this environment. I don't have to buy two or three products and don't have to do some sort of software virtualization of the storage, which takes away your performance.

For hyperscalers, you have insights on the Cisco side that you can use to look at tons of stuff. From the storage side, you could put a range of NetApp tools in the cloud. You could put VMware in the cloud and talk back to the native on-prem Cisco compute environment. 

Their network environment extends into the cloud. There are no limits in terms of integration. NetApp is the most integrated hyper-scale for storage and moving data into the cloud for long-term backup storage. VMware is fully available, so you could run VMware or Kubernetes on VMware in the cloud and tie it to your local storage through NetApp integration.

It's an excellent match because your compute could easily DR to the cloud and be ready to go with all your storage without any modifications necessary because of the native integrations. It integrates well with the hyperscalers. 

Support could be more integrated. For example, you might call NetApp, and they'll determine that VMware is the issue. It would be helpful if they could automatically engage VMware and bring them onto the same call to transfer ticket information and work together. 

Ideally, you can call any vendor potentially causing the problem—Cisco, VMware, or NetApp. They should automatically bring in the other teams as necessary if it isn't. That doesn't happen as smoothly as it should.

I have been using FlexPod for six years. I'm in my seventh year now. 

FlexPod is highly stable at this point. However, there were some growing pains. For instance, when I came on board, they had a bad experience with an ONTAP upgrade that created some unexpected downtime. They have had some challenges with how they wanted to use the entire system.

For instance, the primary use case is virtualized workload. The secondary use case used a fiber channel over ethernet with physical servers. That presented some challenges during certain operations. You'd expect it to be automatic because it wasn't set up correctly. You don't discover the issue until you do routine patching or maintenance work and suddenly lose all your connections to a server. We had some of those growing pains early on, but it has become highly stable after six years. 

We've scaled FlexPod as much as we needed in any direction. We haven't even scratched the surface of what it could do. There are no practical limits on scalability as a design standard.

I rate FlexPod support seven out of 10. It depends on which person you get when you call. Unfortunately, I think that's true of most vendors' technical support. If you don't get the right person at the right level to start with, you're in for a potentially bad experience.

I joined the company after the initial setup. However, it seems complex from what I've seen. The team who deployed it had a reseller set it up for them, but they changed it afterward. They had a different idea of how they wanted it to operate. They tried to make it work in their model instead of adapting to what FlexPod was designed to be. I think they complicated things.

They only use one cloud provider, and FlexPod is minimally connected. Their newer environments are more connected, but backup to the cloud is the only use case they're using for primary functionality, and it's somewhat limited. There was no challenge in doing it, but the overall use case for that is still somewhat limited.

I rate FlexPod nine out of 10. FlexPod allows you to build what you need when you need it. "Flex" is in the name. If you need more compute, you add it and don't need to worry about it. If you don't want to use it in the FlexPod at some point, the compute can be repurposed for something else. The same is true for the network switches. When you need more network capacity, you replace the switch and interfaces or maybe increase the speeds on your network interfaces. Replace the network components with what you want, the capacity you need, etc. The same goes for the storage on the NetApp side.

My advice is to start with what you think you need. As your needs change, put new hardware and performance capabilities in place. You add it as required in the same environment space, so you don't have to change your product line. Your disaster recovery or secondary sites all stay the same. The Cisco and VMware stuff inter-operates. VMware is all over the place, and NetApp ONTAP is ONTAP. It's all the same platform, no matter where you put it in the cloud. 

In Nutanix's environment, they've got two or three. You can have external storage because maybe their internal storage isn't good enough. That external storage platform is something outside of the converged environment. You've got to manage it separately. You've got two hypervisor capabilities here. You could put VMware on it, or you could put Nutanix on it. They're not interchangeable. Once you pick one, you are stuck with that one forever. And if it's not good enough, you have to tear down the whole thing to start over. You're not tied into anything.

We started to move from rack-mounted servers and we needed to make a virtualized environment. One of the requirements for virtualizing all our bare metal infrastructure was to move to a solution with components such as VMware and central storage. We started to look for the environments and were seeking out which was the best version with the possible solution that was in the market and we found NetApp FlexPod, one of the most flexible and easy to use, ready-to-market solutions. We chose NetApp FlexPod due to its flexibility and ease.

The solution is flexible. It's very easy to implement together with the Cisco UTF firewall. We have a computing environment based on the Cisco UTF firewall for computing. The storage we have is the NetApp 3200 series. The virtualized technology is VMware. Together, these three components are very easy and flexible to implement.

I am not familiar with the new technology from NetApp, and therefore am unsure of the latest in terms of FlexPod's native integration with hyper-scalers. Most of the solutions that run now, run on top of the FAS drive or FAC drive. This will improve more and will gain a new level of performance for the new kinds of solutions and technology that are coming out.

We still use FlexPod as a parallel environment. It is a very nice technology. We don't have any pains with this environment yet. That's why we still run this in parallel as we didn't finish the switchover to the new technology.

We use FlexPod's pre-validated architectures. At the time that we designed the solution, it was based on pre-validated architecture, and we had support from the company that we worked with in order to re-validate the solution. With this integration, we needed some support from a specialized technician. Since we used pre-validated architecture, it was simple to improve. We were able to download and implement this solution with no effort. We did this ourselves.

We feel confident that we did something that is custom. The time to market is also fast with pre-validated architecture. We know that if we follow the rules we will get business as soon as possible.

The flexibility, operational efficiency, and scalability of the solution altogether are good. We have two main sites. With this user-friendly environment, we can make both sites replicate each other. When we talk about business continuity, it's easy. We can take the key indicators and our implementation is ready and works as we need it to. There’s also flexibility to scale in. We ran out of capacity after five years and we could scale it in within one or two months and get back to business with confidence.

The solution has helped shift capital and resources to other IT initiatives or projects that had previously taken a backseat due to budget constraints. This is not due to the supplier. Rather, it's due to the kind of organization that we are. We are a nonprofit organization. What can we do is create a government license that provides us with designated suppliers, in this case, NetApp. A special government license can be created with a low price or some other agreement in order to reduce the budget.

The solution helped reduce troubleshooting time on architecture configurations. It's very easy to understand that we follow a pre-validated design when we have good implementation. It's very easy to solve any issues that may arise. We only have to compare what happened before to what happens now and what has changed during that period. Of course, if this is beyond our skills, it's very easy to ask for support to help.

It is difficult to say how much time was saved as we didn't face any outage problems. We didn't face any downtime problems throughout the years. Compared to what we had before, it was not a centralized storage environment. Centralizing changed a lot as we came from a decentralized storage environment to a centralized storage environment and we used a converged technology in this environment. On one technology, it can run on a schedule, it can run cyber channels and it can run any kind of block operation protocols or even file operation protocols for storing the files or the data.

When you are in this kind of environment, you reduce a lot. It's one environment where you can do three or four connections to the storage. Then, you can use any kind of environment with the same solution.

We also reduced our total cost of ownership and simplified operations with the solution's flexible consumption. This is a bundle which is made of three environments, the virtualization and the computing nodes we used with Cisco and the centralized storage with the NetApp, this reduced a lot of space.

It reduced the total cost of ownership. It comes from a different platform and different architecture, and one needs to have more than three or four skills to support their environment. With the bundled environment, we only need one. It's very easy to support this kind of situation.

It would be quite difficult to understand the amount of money saved. As a government organization, we use our partners. Most of the time, when we implement change for new technology, we need to coordinate as people are not adept to change easily. They need to be trained. This is another cost we have to account for and pay for.

With this product, however, we had no difficulty in maintaining the same team. They transferred over from the old environment to the new one. We saved right there.

I ran two data centers. Each data center had no less than one hundred rack-mounted servers. When we consolidated, we reduced our support costs, space costs, and energy consumption costs. Money is saved across all those variables.

The big problem now is that all of the technology is reaching its end of life and we didn't refresh anything at the right moment. Now, we are moving to a new solution. During these 10 years, it was very nice to work with NetApp, Cisco, and VMware together, especially with NetApp storage. We didn't have any problems during this time. I could count only three or four times that we asked for support and this was only to change hard drives that were blocking something. It's been issue-free.

NetApp needs to improve the user interface to make it easier to work in this environment. The older version is poor. However, I'm not sure what they are doing to upgrade the look and feel of the newer version.

NetApp needs to talk to the clients and see what the clients want out of the cloud solutions in order to move more effectively into the cloud environment. It would be ideal if customers could go to a dashboard. They need to sell not only the infrastructure but also the service and both need to be impressive. That's why NetApp should talk to clients as much as possible. The closer they are to them, the more understanding they will have in terms of what a customer wants. 

If the solution offered more workshops and presentations, it could be helpful to lure clients.

I've been using the solution since 2010.

It's quite difficult to understand the tech support in this kind of environment. The three components that make up this bundle that we created in 2010, composed of VMware, Cisco, and NetApp, make it quite difficult. I cannot understand what kind of error it is if I don't understand where it comes from. I need to figure out if this is a VMware, Cisco, or NetApp problem.

I suggest creating a team inside NetApp, Cisco, or maybe VMware, and this team should have the skills to support the companies that support this kind of solution. This will be good as you will reduce the amount of time that you need to solve the problems. Right now, when we call NetApp, NetApp support does not understand what the solution needs and calls Cisco to ask for support. There needs to be some sort of contract or strategy that is better for the client, where the three are integrated together.

That being said, I've never had problems with NetApp, even in these situations. I know a tech professional who was able to guide me through the support process. The contact that I had with NetApp had information that can be found in the web guide. I never had any issues when I needed to get support from NetApp during this period. I've been mostly very happy with them.

Positive

We're transitioning to another solution right now. The main problem is that we don't have support anymore from NetApp due to the fact that the solutions we designed are end-of-life. We need to design a new solution.

The solution is very easy to implement. 

We started with ONTAP, version 7.0. We have NetApp’s 3200 storage series and that is what we use now. It's still version 7.0, with the live firmware.

We are a government company. When we design a new solution, we cannot point to the technology that we want to use. It's against the government's rules. We need to design a general solution with the main points that we want to cover, and the main points that we want to remain. We will sometimes have to choose between several technologies and several offers that we find on the market. That's why most of the time it's difficult to keep the same technology for long.

I'd rate the solution ten out of ten. It is a very flexible solution. Its support, usability, and even the scalability of it has been great.

The way FlexPod has set up our servers has helped our organization. Their OS is on Netapp.

We’ve seen an improvement in application performance. I don't know the percentage off the top of my head, however, after migrating a lot of data from physical servers to virtual servers and putting them on there, it's just amazing.

It increased staff productivity. I was running most of the locations alone. With this solution, I was able to help take care of other problems instead. We ran it and didn't have any problems.

The solution streamlines our IT admin. For the most part, once I get the system set up and put in, adding the VLAN is very easy. Then, users are just adding in VMs. It goes smoothly. I just had to set the solution up and let it run.

Regarding the solution's private, hybrid, and multi-cloud environments, it works well if the communication stays up. The solution’s infrastructure enables us to run demanding and mission-critical workloads.

The solution's ability to manage from edge to core, to cloud, and support their data and computer requirements is pretty good.

The solution is innovative when it comes to computing, storage, and networking. Once it comes together, it's pretty easy to manage. The only action I’ve got to really do is manage everything from my cores or from my distributed switches.

The solution's granular scalability and broad application support helps us meet the needs of diverse workloads. For instance, two years ago, I had a client that was in a building. They had seventeen floors, and I was able to diverse each one of the floors, which were different companies inside FlexPod, and manage it using VCF or VMware. They had their own clusters and it was easy to manage.

This solution is very resilient. For the most part, since about 2018, my servers have had no problems. I still have servers that have been up for years without any issues.

The solution reduces the time required to deploy a new application in some ways. UCS itself is just a hardware platform. VMware is actually where that question is more tied to. 

The solution reduced data center costs. At one of my locations, we had about twenty-four racks full of physical servers. I ended up migrating everything to virtual platforms and putting it inside. It was a decrease of 68% in the total cost of energy. That included the A/C units always running and the power being used for the servers themselves. 

The solution has saved us money. I wouldn't even know how much precisely. I would say $100,000, however, that's likely really low.

I have seventy-six B200 servers. If a server goes down, I have a lot of problems. I’ve also been having random DIMM errors. 

Their DIMMs have been terrible recently. This is a new product that I got about a year ago. My DIMMs are dying left and right and server blades are not being able to function. However, with that being said, when those go down, I have a set of spares that I can put in, and everything works without a hitch, with no problem.

I've had problems with the remote data centers going down due to the connection dropping, and they were not aware that the communication is down. When a link went down previously, the systems didn't know. It then fixed itself. As long as the connections stay up, it works. If the connection fails, it won't.

In my experience with the validated designs, I've always had to go inside and adjust them. I understand that some of them are a base, however, some customers believe that they are 100% proof, and they try to implement it. Then I get called in to correct the errors and correct some of the layouts to include some of the newer features.

Since 2018 or 2019, maybe due to COVID and the chipsets, my DIMMs are dying left and right. That's the only problem I have. My boards are fine. The servers are working fine. 

A feature I would like to have in the next release is an application desktop that talks to it, so I don't have to go to the web GUI as much. Besides that,  it's pretty bulletproof. I use UCS over HP and Dell, nine times out of ten.

I'd like to see a little bit more versatility with the C220s and the C240s, to see the expansion ports on those servers grow. Besides that, everything has been pretty amazing.

I've been using the solution since 2012.

In the past, stability was 100%. Recently, it's been terrible due to the DIMMs.

It's greatly scalable. It just relies on what FIs you have for your interconnects.

In terms of support, I call in when there's a bug. I’ve had problems with the memory as well. I had a server that was DOA, and it came down to the fact that we didn't even know what the problem was. That took almost a year to resolve. Then, with the DIMMs, it's taken me about two months due to the testing they want to do. That said, in 2018 it was phenomenal.

Technical support needs a little bit of work when it comes to hardware. In terms of software, they're not too bad.

Neutral

I used HP and Dell. I was having a lot of problems with HP and Dell was getting expensive. I had a little extra cash to buy the UCS when it first came out. There was a little bit of a learning curve, however, once I got that down, it worked well. I'm a big supporter of Cisco.

Currently, we're using the 4.2.1 with M5 servers, B200 M5s. This is my third one, as far as the firmware updates, and driver push-outs.

The initial requirements for getting it up and running are very complex. There's a lot of note keying, and all of that had to come into play. We had to have a good foundation of server networking. It's not something anyone can just throw a user into and say, "Here's a gear," and set it up.

We have seen an ROI.

In terms of the cost, the last bill I saw was about $3.5 million from the latest contract. That might have been for a five-year contract. That was with the licensing for the ports for the FIs, along with the tax support and the software assurance with Cisco tech.

We have not used the solution to integrate advanced cloud services. I’m working on a VCF project currently. I have not used Intersight, Active IQ, or CSA yet. That's actually on my to-do list for my current project.

At this point in time, we do not use this solution to power any AI machine learning applications.

UCS is more network-driven than it is server-driven, which is what Dell and HP, drive on. Once we set up the basic server parts, the rest of it is network base. It is a mind changer. When I handed it off to server admins, they were worried about a lot of issues that they used to deal with on a Dell, HP, or even IBM. They don't have to worry about that with the UCS.

I'd advise new users to understand where they’re putting their ports and know which ones are going to be fiber channel ports, FIs, and make sure they have a distributed switch and are not connected directly to the core switches, the 7Ks or 9Ks. I've seen people take down their whole environment when somebody added a VLAN or added a network for the UCS network that was already on the core. It took the whole thing down.

I'd rate the solution ten out of ten. If the DIMM problems were removed, by far the solution deserves a perfect ten.

We have high-IO SQL workloads. We have over 200 SQL Servers for our flagship platform so we definitely need compute, storage, network, all of that in one, that's going to perform under pressure.

We're using AFF A700s, which at the time we got them, were the latest and greatest of all-flash storage. The Cisco UCS portion that we're using are M5s and M4s. They were top of the line when they were released, as well. 

We're looking for low-latency and high-compute, and that's what FlexPod gives us.

We have an overnight production run. We take a bunch of files from some of our clients and mash them around and throw them into the databases, and then processes them in SQL. We've been able to reduce that run-time, just by upgrading the UCS portion, by 20 percent in the past year.

We've implemented a lot of initiatives over the past five years, but bringing in SSD was the big one. Then we added more controllers and updated UCS hardware. Those are all steps that have enhanced our application performance. This year, we also adopted SnapCenter, which is a NetApp product, and that has increased the reliability and efficiency of our backups as well.

UCS has also reduced our data center costs. We had HP machines, which took up the better part of 2 racks. Bringing all that into a UCS chassis, with the FlexPod solution, has reduced power use and it has reduced physical footprint and has cleaned up our racks. We have easily saved over 50 percent physical space. 

One of the most valuable features is the interoperability between the devices, between the Cisco, NetApp, and VMware. That's always nice. 

The supportability is also good, the fact that we can call one vendor and they'll help us. We don't need to call our vendor, Softchoice. We could call NetApp and/or Cisco and/or VMware, and they would all help us. We wouldn't be pushed away to the sides. They're not going around blaming people. The solution is sold as-is and it's supported by the three parties. They have to support it, and that's nice.

The solution’s validated designs for major enterprise apps are also extremely important in our organization. We don't always have time to research products and solutions ourselves. By going with a validated design, we're assured that it's the latest and greatest. It's supported by the three major vendors that we deal with. That's not really something we could find with other vendors, although, to be fair, we haven't looked around.

I'd like them to bring back the GUI for NetApp ONTAP. They changed the interface in version 9.8, and it's not great. In 9.9 they've tried to make it better, but its still as useful as 9.7 and before. 

We've been using FlexPod for about eight years.

It's extremely stable.

The way that we have NetApp built, with HA, it provides redundant workloads. The storage failover is pretty transparent, so when there is an outage, none of our workloads is affected. It's proven and tested. They throw the term "non-disruptive" around a lot, and it actually is non-disruptive. Obviously, I was hesitant when I read that, and I wanted to test it for myself. But I've personally been involved in some of the storage work that's been done over the past two years, and I can agree that when NetApp says it's non-disruptive, it is, in fact, not-disruptive.

Over the years we've expanded our FlexPod, our NetApp nodes, from four to six.

It's scalable. If we need capacity, if we need to scale up we can purchase new blades and have more powerful CPU and RAM. If we need to scale out and have more space, it's pretty flexible. Whether we need to do storage, compute, or network, they are all components that we can just purchase and hook in.

NetApp has definitely been there for us. They're a good partner of ours. We also use our third-party vendor called Softchoice. They're our primary support guys and we go to them first, and then they will open a ticket with Cisco or NetApp or VMware, if necessary. When it comes to NetApp, when we have needed help their support has helped us. The unified support for the entire stack is extremely important. The fact that we can just call one vendor and get support on it is a huge bonus.

Positive

NetApp has been with our organization since before I started working here, although back then it wouldn't have been a FlexPod solution. It would have been a piecemeal solution of HP and NetApp.

It's expensive, but when you want entreprise services and support you pay for what you get. 

We still require the proper skills and the proper people in place to manage it. It's still a network. It's still storage. It's still virtualization. I wouldn't consider it a small-office type of solution. It's definitely a data center or enterprise-level solution. But it is a little bit simpler than if we were to piece together the solution ourselves with other vendors. If we were to get an HP and build ourselves a NAS, for example, or even if we would get something that's not supported the way that our FlexPod solution is, it could be more complicated.

I don't think the solution has saved our organization in terms of capital expenditures because we do upgrades, either because we need space or because we need compute, every year. But I wouldn't say that's a bad thing either. It's not like we have drastic spending. It's a matter of trending. If the business is doing well and the application and the platform are doing well because we're onboarding more clients, we need more compute and storage.

I have been working with FlexPod for a while now. I recently shifted my job and have been working with a solution included in FlexPod. Most customer use cases that I have seen are either using it as a database management system or for a VDI solution.

There are a lot of points for configuration.

We are using a private cloud with Azure, but the newer versions integrate with Cisco Intersight.

You get data privacy with it. 

The solution helps to optimize our operations with insight gained from Intersight Active IQ or CSA.

The integration part of things is the most valuable feature. You are getting a whole set of things under one roof and rack. There is support for everything, which is one of the cool things.

The designs are pretty good. Cisco, NetApp, or the OS vendor keep on updating them, which is one of the good points. They will send out a new document about a design refreshment. Everything integrates perfectly with Cisco's new chassis and NetApp version 9.9.

The different modules perfectly integrate with each other because of the Cisco UCS part. For a single chassis, you might have eight plates powering up. Then, there is Nexus, which integrates with your FIS pretty smoothly. For the storage part of it, some solutions have MDSS, and some don't. However, getting them configured is pretty much a few clicks.

I like the continuous CI/CD upgrade cycle with this solution.

They just announced that they are going to move it along with Intersight from Cisco. That can be a private or public cloud, which is one of the areas where it can grow more and has a lot of potential.

I have been using it for somewhere around three to four years.

The solution is resilient.

It has become easier to monitor and automate processes using the solution.

We get everything under one roof instead of just modulating parts.

It is scalable. I have seen the solution used on multi-site environments. I have also seen somewhere around 2,000 to 2,500 people using it on a single site. In other use cases, I have seen it being used in smaller environments, where the data capacity is assigned. Something that I discovered myself, the data relevancy needs to be really good.

If the network or site is down, you just need to go to a single vendor. You don't have to open up multiple cases with each vendor to get things done. That is one of the financial benefits of this solution.

The technical support is pretty good. Rather than running to different vendors, you can open up a case with any of the vendors, who will then communicate with each other to get things resolved. So, customers can go to different vendors for a single issue. From my perspective, if a case is being opened with Cisco, I have seen their people working with VMware to get things resolved. 

I would rate the customer support somewhere between 7.5 and 8 out of 10.

Positive

I have not previously used another solution.

If you use the design document, everything is pretty straightforward. The racking and stacking are pretty easy, in regards to the physical stuff. Cisco and ONTAP are pretty simple to configure if you follow the proper design.

You just need to do a couple of clicks for your UCS. The same goes for Nexus. It depends upon the configuration, but it is pretty easy to deploy. Once that is done, it is just how you want to use your storage, which is the only contribution that you need to do because everything else is taken care of. 

It takes a maximum of two or three people to deploy the solution, e.g., someone to do the physical work and another person to configure everything. 

Once the physical work is done, the configuration part comes in. That is when your switches and UCS integrate with each other. I have done the configuration on Nexus and UCS parts, where I definitely needed help.

We have seen ROI through IOPS and network latency. 

I did not really evaluate other options before choosing Flexpod because it is a leading product in the market for converged use cases.

The private cloud environment is one of the major selling points for it.

Usually, people move to a different solution when it comes to getting a hybrid cloud solution, e.g., a CA solution or HyperFlex. This is where I have seen it get a bit distorted.

I would highly recommend it for core and multi-cloud solutions.

The way that they are making the progress, it will still be a relevant solution going forward. Where there is a need for big data, this solution can be considered.

I would rate this solution around 7.6 out of 10.

We have FlexPod Mini for the primary data center.

FlexPod's validated designs for major enterprise apps in our company are important because there is stability. There are zero downtimes and high availability. There is good support for the systems that you can run on the platform. FlexPod is a validated architecture, and basically, the spectrum of what's supported is pretty wide. So, you can run pretty much everything without thinking twice about it.

It provides unified support for the entire stack. For example, if you have an upgrade or a new version on NetApp, there is a compatible version for the Nexus switch, and there is a compatible version of VMware and/or Cisco UCS firmware. Instead of upgrading piece by piece or guessing what is going to work with what and whether there are any bugs, for an upgrade, you can follow the chain and what has actually been validated. It reduces a lot of overhead for the team.

It has made our staff more efficient, enabling them to spend time on tasks that drive our business forward. Instead of designing or trying to follow the lifecycle of each piece of equipment, by working with a unified stack, we do it once, instead of doing it five times for five different pieces.

It has definitely improved application performance in our company, but I don't have a baseline.

Integration is most valuable. This is a reference architecture. So, we don't have to design something from scratch and figure out how it is going to work. 

We would like one-click upgrades.

NetApp released a new version with a new interface. For somebody who has been used to the old interface, it's a change. It is taking time to adjust to the new interface, and it would be nice to have some of the old features in it.

We are very positive about it. It has been a great experience. We've actually refreshed the hardware which indicates that it is working and is stable. We are satisfied with it, and we're just continuing with this.

Our experience is positive. We've refreshed it. We've purchased additional NetApp, which speaks of the positive experience. I would rate it a nine out of 10.

This was our first experience with it. Before this, we'd buy hardware, storage products, and networking products, and we tried to integrate them. Whatever surprises we got, we dealt with them. With a validated architecture, there's a little bit more confidence that whatever you're putting in place has been validated, and then you got two major names, NetApp and Cisco, behind you.

In technology, I'm afraid there's really not much that's straightforward.

We have some skills to do some of the tasks, but for implementations, we usually go for integrators. The experience with the integrator was great, and the time was basically within an acceptable timeline. The project timeline did not extend, and from that perspective, the implementation was straightforward. You can have some expectations for start and finish.

We worked with our integrators to look at the available solutions and follow the market trend based on our requirements, and this one checked most of the boxes. At the time, instead of NetApp, there was HP storage or HP servers with HP storage. Based on the previous experience and experience with the staff, integrator's feedback, and market popularity, the choice was Cisco/NetApp.

If anyone is just going from a conventional SAN to VMware Hypervisor, it is the most reliable option moving forward. Following technology trends, if you're moving from a conventional server to SAN and you would like to integrate from encryption to SAN-to-SAN replication to any features—ranging from security, ransomware protection, and DR—this solution covers it.

It simplifies infrastructure from edge to core, but I don't know if it also simplifies from core to cloud. 

We are not yet using FlexPod's storage tiering to a public cloud. We also haven't fully adopted most of the innovations, such as all-flash CI, private and hybrid cloud deployment, secure-multi-tenancy, end-to-end NVMe, cloud storage tiering, but we are getting there in terms of whatever trends are there in the market within cloud integration, flash, and NVMe. It is improving our infrastructure, and we will be there. We are currently in the process of adopting some of these.

It has only theoretically decreased our company's data center costs in terms of floor space, power, or cooling. That's because when we went into FlexPod in a data center, we were migrating from one data center to another. At the moment, they still coexist. We are still in transition. So, in terms of cooling and power, we are still cooling and consuming power in both locations. Until we completely go off one of the data centers and move some of the workloads to the cloud, practically, there won't be any reduction in the data center costs. 

I would rate it a nine out of 10.

We're using it for general purpose virtualization or converged, as well as in specific cases like electronic medical records. That is the big one.

In the partner space, it gives us a validated solution that we can deploy and it's very repeatable for us. It helps our customers in that they can have confidence that it's going to work exactly as it's supposed to.

It has also helped reduce troubleshooting time—easily hours per week—on architecture configs.

FlexPod’s prevalidated architectures are very important to our organization. It has to do with predictability for applications that are always up and that sometimes are life-safety or life-critical applications. Especially in healthcare, it is absolutely critical that we have a validated performance platform. It has to work every time.

A lot of small things could be improved. I'd like to see better integrations with some of the third-party tools, like Terraform. That would be good. We use Ansible to deploy and that's good, but it's slower than it needs to be.

I've been using FlexPod for more than five years.

The stability is a 10 out of 10.

We haven't done much scaling yet on this most recent one, but in general, the scalability is very good. It's a 10 out of 10. It's very easy to grow very big.

The technical support is good. It's not perfect, things never are, but we've had very few issues. It's also relatively new. We'll see in a year. Maybe my opinion of it will go down, but it's been good so far.

Positive

I have experience with Vblock, Vxblock, and FlashStack.

With FlexPod, we have a lot of validation around performance. Especially in the medical world, it's a very well-known entity, so we don't have to struggle a lot with finger-pointing. Those are all good reasons why we picked it.

It is a complex deployment, but we have done it a lot of times so it's not that hard. We have it all scripted. We have a ton of automation in the deployment process.

For healthcare, it is almost always on private cloud. That is still very much the standard. It's mostly Azure and some AWS, a little bit of GCP, and some others. One of the big EMR providers has its own hybrid cloud that is purpose-built.

The most recent one I did was a big EMR. It's a moderately sized NetApp AF series and a bunch of Cisco UCS with NDS storage. It is a reference flash tag straight out of the CBD with 150 nodes.

Our customers definitely see ROI. We generally model the TCO for them over time and we're generally pretty accurate. They usually get their payback on the product-based converged solution in two years or less. They usually avoid having to add headcount.

The solution's flexible consumption has definitely reduced our customers' TCO. It allows them to do more without their having to add staff to support it. The flexible consumption is a good option for some customers and not for others. We have some who love it and some that don't.

They're going to spend the money on the solution one way or the other, and flexible consumption lets them spread it out over time and pay as they grow. That's great for some, while others just want to do the CapEx because of tax reasons or the like. Neither one is better. They're just different and they're both fine.

Overall, the solution works pretty well. The biggest complaint I have from customers is the cost.

The flexibility, operational efficiency, and scalability of FlexPod are very good. We also use other products too, like FlashStack, and these solutions are equally good or similar in most ways. I have a very good opinion of FlexPod, and we've been using it for a long time.

In terms of comparing converged infrastructure solutions and picking the most cost-effective one, you have to pick what works for you. Think about who's going to support it. If you're hiring a vendor, like me, you want to make sure that you trust me and that I'm going to be around. If you're doing it in-house, make sure that you're picking the one that your people can run.