FlexPod Review

The agility reduces the number of hours that it takes to construct a physical or virtual data center

What is our primary use case?

The purpose of FlexPod is for a converged infrastructure that provides compute or networking storage and helps launch applications more easily and dynamically.

How has it helped my organization?

At the end of the day, AI is not AI without the application that we write into it. With collaboration between Microsoft — utilizing it to build in a manner that is compatible to the FlexPod architecture — we're able to provide specific intelligence that supports our objectives — whatever it is at a given time. Whether it's data aggregation, learning, pouring out the analytics, the intelligence helps specific applications respond to requirements within a business structure. That's what FlexPod enables us to do. That agility reduces the number of hours that it takes to construct a data center, whether it is physical or virtual, by enabling applications to support AI objectives. It just needs to be built correctly.

We have experienced about 28 to 30% improvement in application performance and in our industry that's actually a very significant improvement.

The purpose of using FlexPod, for us, is to simplify and streamline application deployment. 
Compared to utilizing a rack and stack model and using a virtualization technology like VMware, the time savings is about 40% in getting the application into production.

What is most valuable?

Certification from both manufacturers states that this is a tried and true converged product. That's what we are most happy about. One of the biggest things that my engineers have the pains with is to vet out core networking, vet out stretch routing, vet out applications and then vet out the compute, the front end and the stores, then layer it. After all that deal with the application and quality assure it before we put into production. FlexPod cut out all that complexity and helped get us to the point where it in a data center, launch our application, build the application, test it, QA, and then put in production. So it does reduce the time with regards to how we dynamically provision and provide applications to our end users and developers. 

What needs improvement?

If we look at data center solutions, any of those solutions are only as good as the people that put it together. If there's a way for us to take a hyperconverged technology or converged technology — like FlexPod — and use it with artificial intelligence, that allows the engineer who is building it to infuse the deployment with intelligence. Turning it on, the necessary steps — done correctly — eliminate human error. If something is in error or not within compliance to confines of what that particular architecture should be like, intelligence lets that engineer know that an object is out of policy. For example, if you implement SAP and Oracle, the Oracle database goes through this way; if you partition it out to this number of lines or a particular number of virtual machines, the recommendation may be different to achieve the maximum efficiency.

If the solution does that, it helps enable and accelerate deployment. Every organization out there has its own challenges. Whether you're an automobile manufacturer, or a cloud solution provider, or a managed service provider, or even application software provider working for social networking where the only thing they need to do is support people, all that is important is when they login to that particular application. They need to have that effort fit the user experience. The collaboration between Cisco and NetApp can learn to provide that.

Millennials today are very intelligent people when it comes to social media, but they're not hands-on with applications or as CLI (command-line interface) as some of the older engineers. The millennial comes in and they look at something and they get it. Okay, as long as that's valid, it is okay. The smarter thing is that something is put into FlexPod to be sure potential errors are covered. 

The client will tell you what they want to do. Well, whatever that is — they can be selling hamburgers, make pizzas, or fly an airplane. If we make a machine dynamic, it allows professionals to go to market and set their strategy a lot better.

What do I think about the stability of the solution?

As far as stability, the product is a tank. It doesn't break. It's very reliable. It is also resilient in terms of workloads, but it has to involve the necessary security staff to oversee it and the proper security application and layers to support it. But structurally and architecturally, the solution itself, from a workload or a workforce perspective, is very resilient.

What do I think about the scalability of the solution?

It does have its limitation if the architecture is weak and constructed incorrectly. If you do it right, it scales infinitely.

When you build it, and you build it to scale, you'll be able to serve out any application dynamically to end users. It could be an organization of 3,000, it could be an organization of 50,000. As long as you build your FlexPod architecture correctly within your data center, whether it's a co-location or a physical data center, it's proven itself to be extremely scalable.

It becomes an Achilles tendon when an organization leveraging FlexPod does not build enough scalable resources. That's when layering applications does cause issues. I've seen that both from a security perspective, as well as an application performance perspective.

How are customer service and technical support?

We use technical support all the time. The collaboration between Cisco and NetApp is actually very good. We use both platforms. Even though we work with Cisco directly to utilize HyperFlex architecture, which competes with FlexPod, the customer service isn't competitive and remains collaborative. There is no finger-pointing, which is very surprising. More often than not, we're able to satisfy an anomaly or technical issue easily. The technical support is very, very good.

Which solution did I use previously and why did I switch?

We hated taking racks down or putting them up just to deploy a simple solution. If we need an application and had to put another rack up, it means using a lot of resources. Instead, we could launch a virtual machine. The network, the compute and the storage is in a single solution.

If you have to spend more time during a day fixing computers, servers and the network than you do focusing on what you make money from, you don't need to be in the business you are in. That's why they provide hyperconverged technologies that are data-center-centric out of the box. You buy it, you bracket, turn it on, load an application onto it, and then you build it. 

It all started many years ago when IBM created the most intelligent compute system in the world. Everybody logged into a VT100 terminal. They didn't care about what was going on in the machine. They logged in and it worked. Then some guy decided to break it apart and create a disparate network. When they figured out they realized it was too sophisticated. As the company grew they needed a server for every single application. That's why you see the evolution of VMware and Citrix and the evolution of converge.

The future of things moved away from just hardware. The future of things now is going to be like hyperconverged but in a very virtual form. That's the reason why Cisco is buying organizations like BroadSoft. They want to get into organizations that provide virtual services.

How was the initial setup?

The product is actually easy to set up. It's self-learning. It's methodical. At the same time, you have to go through all the minutia of the networking layer, the storage layer, the compute layer to focus on the foundation. Then prepare it for application download and then application build on either databases or the application itself based on the OS that it resides on. The model is quite simple.

What about the implementation team?

We do the implementations.

What was our ROI?

People go to the cloud today and think that it's going to save them money. Actually, if you're going to go to the cloud, you're going to spend more money. The difference between going to a cloud infrastructure and having your own private cloud like say FlexPod, the cost structure is the same. You're going to need humans to continue to manage, maintain and run it. You're going to continue to do a refresh on it because technology will get old. Cisco and NetApp will never sit on their laurels. They won't just create FlexPod and have only one model. Over time, switches, routers, storage, interfaces and things like that will change.

That's why I think it's important that we don't focus too much on ROI. ROI is not the amount of money you spend on FlexPod or cloud that equates to revenue. ROI is whether you have a good product that allows your company to leverage technology. FlexPod enhances the way you go to market. That is an ROI. 

If a CFO wants to do a 10-year map to see how long is it going to make up the investment, you don't need to buy FlexPod. You need to talk to how you to go to the market efficiently. You needed to ask yourself whether your company will be viable and competitive to stay in the market landscape with respect to what you sell. 

You have to understand why you're spending that money. If not then this investment will not make sense.

Which other solutions did I evaluate?

We considered VMware, Citrix, going full cloud, sharing with a cloud, handing it off to a managed service provider, building it ourselves, rack and stack — pretty much everything was on the table. FlexPod is a good product. I think they just need to continue to keep up the pace with organization like Nutanix and those types of organizations to be able to compete.

You can't get in trouble going with Cisco and NetApp. If you get stuck or have an issue, support is there. The inner partnership, inner engineering, and cross-pollination is there. I'm still leery of some of the up-and-coming hyperconverged organizations out there trying to compete. They may be good, dynamic, fast, growing, everybody's getting on on it, but they're not backed by two large publicly-traded organizations that have a legacy foundation that's been tried and proven for what they do and do best. 

What other advice do I have?

I would probably give this solution a seven-and-a-half or an eight out of ten. It isn't higher because I know that if I were to look at a very dynamic data-center solution, there are organizations who can do it a lot more agile, more quickly, or in a more user-friendly way. It takes a very sophisticated group of people to run and maintain NetApp and Cisco. It's not just a box you put in a server. You scale it out and you log onto a graphical user interface and you manage it. When it is running, it's a very, very powerful foundation that no other hyperconverged solution out there can compete with. You cannot break it. And like I said, as long as you have the right people who know the foundations, FlexPod is a very powerful data center foundation.

I think one of the greatest things that we like about NetApp is the fabric OS and leverage that proprietary app to be able to make it self-aware of legacy storage, legacy compute, current compute and future compute.

One of the cumbersome parts that we discovered is that there are claims that say something can be done, but it takes a lot of testing and trial and error and working with our ISP to ensure that these multi-cloud, multi tendencies and applications living in it all talk to each other. In other words, it's not going to run by itself. It will continue to take a group of highly sophisticated engineers and application folks to be able to make things work.

FlexPod was built in collaboration with Cisco when they didn't have their own hyperconverged technology and when NetApp didn't have their own networking technology. The idea behind FlexPod was to build that converged and hyperconverged foundation to support it. The direction Cisco is moving in today leaves the partnership intact on that app for now, but with some of their hyperconverged solutions out there it may not stay that way. Competing HyperFlex technologies are extremely agile today, and if they continue to develop, possible partnerships with the likes of Oracle or Linux or Microsoft may be something to be reckoned with.

There are no walls to technology. As long as you code out a certified solution to dynamically support your market strategy, that's all you needed. That's what I really learned from blind spots, and that's the reason why we moved in the direction that we did.

Don't look at the price. It is more important to understand where your company is competitively in the market. If you're going FlexPod, it's going to be a journey and that FlexPod isn't going to make you money. But it's going to help you really find your company, or the next level, or the future of where you're going to be in terms of going into a market. You should not buy FlexPod because you want to be cool like other companies. It won't save you money. It is more important that it enables your organization to be more visionary and more technically dynamic.

Disclosure: I am a real user, and this review is based on my own experience and opinions.
Add a Comment