Getting Started with VDI
Different hardware and software components must come together in a coherent way to create a good desktop virtualization experience. Important components, such as GPUs for high fidelity video are often invisible to the end user. But government agencies that are assembling VDI solutions must take all of these pieces into consideration and understand how they depend on each other.
When choosing and architecting a VDI solution, the first question to ask is: what is the workload you need to support? Agencies must not only consider which applications users need but also the kind of data they are accessing. Virtual machines come in preset sizes, so can choose the best size that fits the workload. You also need the right hardware stack to support that VDI solution and optimize the users’ experience.
Supporting the Edge
One consideration for government agencies is supporting the devices at the edge—which have particular concerns with data sovereignty, latency, and accreditation. You also want a solution that can be connected and or disconnected with an air gap as needed so you can support mission workloads and security needs. Agencies should look for the capacity to extend AI to edge devices so you can run advanced analytics closer to where that data is being generated. This allows real time insights into workloads so users can envision new possibilities.
When you use graphics or video, frame rate and the quality of those frames become very important. The workload determines what capacity you need. Are you using Zoom or Teams? Do you have big graphics-heavy applications like Google Earth? Or are employees using large workstation applications like ANSYS, or MATLAB ArcGIS, which require more hardware? Some use cases—such as a doctor looking at an MRI image—require the highest resolution. 30G is the standard within media entertainment because that gives the best image quality. But images are often created on the server and then encrypted, compressed, and decompressed—processes that can impact the user experience.
Hardware solution architects can help agencies understand the workload and put together the right package to achieve your goals. Requirements for use and quality impact your choices at the server and the client side as well as hardware choices in the datacenter; your CPU, GPU, RAM, and IOE all need to work together seamlessly.
It is important for your endpoint devices to be capable of decoding the data stream from the server and keeping up with the content. If you have four times the number of pixels, it will tax the monitor hardware and require the endpoint device to decode more. The datacenter architecting should not be undermined by insufficient endpoints.
Hardware plays a critical role in VDI and hardware flexibility can profoundly affect the user experience. You want a solution that spans everything from the edge data center out to the cloud—whether it’s on premises, in a public facility, or in a dedicated sovereign cloud. Agencies should look for hardware that can work across the entire range of hardware requirements—including compute, storage, and networking. Solutions should offer managed services regardless of your cloud provider.
Government agencies can benefit from hardware as a service, which allows you to divest big upfront costs and distribute them over time. You should also seek out robust worldwide services across advisory, professional services, cloud consulting, etc.
Network is an absolutely critical part of delivering a remote solution. State-of-the-art VDI machines and architecture can be slowed significantly by a bad network. A common challenge is getting data from the desktop from the server out to the endpoint, which is obviously a very network dependent process. The top priority is getting the right media to support that latency requirement that you want. Then you can leverage in-memory caching capabilities to take it to the next level and lower the latency further.
The Role of the GPU
A GPU is more than just something to generate pixels. It provides high frame rates, lowers latency, and allows a remote user experience to feel local. Government agencies always want to reduce network latency and encoding on the GPU is one way to accomplish this. Without a GPU, the CPU must do all the graphics work, which requires compression and slows the process. Having the GPU available for graphics work frees up your CPU to run applications.
You need the right GPU to ensure that there are no bottlenecks in the system. It’s critical to get the right number of pixels and refresh rate at the endpoint. There are a number of services that will bring together all the components into different cohesive solutions, creating a turnkey system that can start working immediately.
View our presentation to learn more about how virtualization infrastructure can benefit your federal agency.