3D FlexPod: The Best of NetApp, Cisco & NVIDIA in a Single Stack

By Kenneth Westerby, Sales Rep, NetApp Norway

 

3D FlexPodLast week, I received a lot of attention on social media (more than 50,000 views) when I posted about how to run oil and gas applications on LinkedIn. Now maybe it was the beautiful dual-screens that grabbed everyone’s attention, or maybe it was the fact that we were replacing expensive workstations with something more simple, more secure, and frankly, much more exciting!

 

What you are seeing in those beautiful screens is a full 3D experience, delivered remotely and with zero performance degradation via a 3D FlexPod. In short, this 3D FlexPod is a datacenter stack that combines the best from NetApp, Cisco and NVIDIA to run workstation workloads as virtual machines.

 

Nowadays, workstations suffer from the same issue as servers did in the early 2000s -- they are often underutilized, and upgrading and maintaining them takes time and money. With the NVIDIA Tesla M60, the ability to give workstations the benefits of running as virtual machines opens up an incredible new way for people who deal with high-end applications -- to access them, collaborate with others, and work from almost any location in the world. This means no more installing expensive workstations in remote locations, and no matter what the device, it gets the full 3D experience.

 

The opportunity goes beyond the oil and gas industry, although that was the purpose of this particular demo. Whether it is automotive, aerospace, or countless other industries, I can think of countless collaborative teams that could benefit from the use of a 3D FlexPod. Just imagine, enterprises could now have all of these remote teams working together on the same data and the same assemblies. That means no more moving data backward and forward! You can even think of this as “Design as a Service,” taking collaborative design to a Cloud model.

 

The 3D FlexPod uses NetApp’s All Flash FAS systems (AFF). The optimization of ONTAP 9, our storage operating system, takes full advantage of the use of flash to provide ultra-low latency so time consuming applications simply run better. This combined with the M60/M10 from NVIDIA makes sure that applications run at optimal speeds, thanks to the VM being on an operating system designed for low-latency.

 

Of course, storage is just one aspect of this solution design, but it’s an important one. Working on large data models and assemblies requires significant storage IO. The lower the latency, the better the experience for the user. In many collaborative design environments there is typically a Product Lifecycle Management system backed by one of the leading database vendors, so the storage device must be appropriate to support and protect this as well. But not all are. It’s these high-throughput, low-latency requirements where NetApp solutions come into their own. And of course there is the proven design and support for the joint NetApp / Cisco FlexPod solution

 

VM’s can run better and you have all the benefits that a virtual stack can provide you in terms of end user CPU, Memory and OS requirements. But where this really gets interesting is when we add two more features two the stack.

 

The NetApp Data Fabric with ONTAP OnCommand Cloud Manager

A workstation has very little benefit from cloud computing, but VM’s do. Imagine you are based in Europe and need to support a project in another part of the world. That is a well-known challenge and one that takes a lot of time and effort if you are dealing with physical machines.

 

In a 3D FlexPod environment, things are different. Using Data Fabric, you can choose from hundreds of service providers around the world and simply have your CPU and GPU on demand. Find the service provider that offers you the locality and performance you need and simply use OnCommand Cloud Manager to extend your reach. You can have VM’s running in your own secure hybrid cloud environment using out-of-the-box technology that comes with NetApp ONTAP 9.

 

Collaboration

Since we keep all the data, CPU and GPU’s in a central repository, all that ever leaves the data center is pixels. You can use a 40GB network to load your data into the VM and from there you can access the application consuming only the bandwidth it takes to move the pixels. This is normally less than 10mbits. The most important part is that now we are no longer moving files, just the graphical display. More than one user can share the same desktop and discuss an image in real-time. At VMWorld Europe in Barcelona, we ran this demo showing oil and gas, CAD and medical applications running in real-time from Oslo.

 

 

In conclusion, FlexPod is already leader in converged infrastructure, but adding NVIDIA element and making it a 3D Flexpod now allows you to run applications that require Quadro Graphics. This means you now get all the benefits of the combined NetApp & Cisco collaboration and all the tools you need to build a new type of virtual infrastructure that all revolves around how people turn data into information.

 

Oh, and did we mention you can run 3 x 4K display with ease?