Hyper-converged infrastructure systems (HCI) come with an almost bewildering array of options. While it’s possible to select a HCI solution that is generally tailored to host VDI workloads, you still might need something more customized.
Matching HCI technologies to your VDI infrastructure is more complicated than you might think, due to the variety of ways that organizations deploy VDI. You can use the presentation virtualization method, where users establish remote desktop sessions on a server. The server hosts separate user sessions that appear to be full desktops locally, but dont require using virtual machines through a VDI client. Then there’s the RemoteApp scenario, where an application runs on the server without using virtual machines, but only the application output, rather than the full desktop image, presents on the client. In both cases, the HCI solution doesn’t require optimization for VMs. There’s also the more traditional VM-based approach, where users connect to a VM running on a server. In that case, you would select a HCI deployment that is optimized for virtual machine hosting.
To select the optimal hyper-converged infrastructure solution for a VDI workload, it's important to understand the workload characteristics. You need to know which computer resources the VDI workload is most likely to use: memory, graphics and processor capacity as in the case of computer-aided design workloads? Is it storage, network and processor capacity as in the case of video-editing workloads?
This knowledge allows you to determine which resources should take priority. While it would be great to simply max out processor, RAM, storage and network capacity whenever you wish, it’s extremely costly to do so and likely unnecessary. Instead, determine which hardware resources will likely cause bottlenecks for running the workload and then ensure that the HCI configuration is provisioned to minimize that risk. For example, ensure that you have adequate processor capacity for workloads requiring extensive calculations.
The tighter the alignment between the hardware configuration and the workload profile, the more efficiently and quickly your VDI sessions will run.
Underwritten by HPE, NVIDIA and VMware.