This section provides various general best practices for sizing your VDI deployment.
Platform configurations
You can select the instance density of the subscription with three configuration options (16, 32, 64 instances per host) for Compute, General Purpose, Memory and Large-Scale Memory-optimized instance types. This enables you to be more in control of the number of hosts provided in each instance quantity selection.
For example, if you have an availability requirement for Failures to Tolerate (FTT)=2 with RAID6 (minimum of six hosts), you can choose the “16 instances per host” selection to ensure you get at least six hosts to meet your requirement without paying for unnecessary instances.
For VDI, the "Memory Optimized" instance is a good fit when you are looking for higher VDI desktop density. This configuration provides a good balance of performance and scalability for various general-purpose VDI workloads.
The "VDI Optimized" instance provides the highest levels of performance for more specialized VDI workloads requiring GPU acceleration for graphics.
When using the “Memory Optimized” instance type, use the following sizing recommendations:
- Knowledge Workers—Up to 2.65 users per instance, for example, 510 Knowledge Workers would require 192 General-purpose instances.
When using the “VDI Optimized” instance type, use the following sizing recommendations:
- vGPU Enabled Knowledge / Power Workers—Up to 2 users per instance, for example, 128 vGPU-enabled Knowledge or Power Workers would require 64 VDI optimized instances.
The following figure shows instance types:
CPU
Choose the APEX Private Cloud or APEX Hybrid Cloud physical instances to specify the CPU core count of the host servers. Select 16, 32 or 64 instances per host to determine the number of CPU cores installed in your host.
Memory
Choose an APEX Private Cloud or APEX Hybrid Cloud instance with a high enough memory to core ratio that suits your VDI workloads needs. Instances range from small (4 GB) to extra-large (32 GB) memory-to-CPU core ratios.
Storage
Depending on the Instance type selected, you will be presented with three storage capacity options to choose from for the first cluster you configure. This will be presented as Raw Terabytes (this does not account for RAID levels or slack space requirements) and will be based on the quantity and type of instances you selected earlier.
Because Dell requires uniform configurations within a subscription, if you choose to configure additional clusters, you will only be presented with a single capacity point for those subsequent clusters as well as for any capacity expansions you may add over time. For this reason, we recommend you select the larger capacity point for the first cluster to ensure you have sufficient storage for additional clusters and expansions.
NVIDIA vGPU considerations
Choose the VDI-optimized APEX Private Cloud and APEX Hybrid Cloud instance GPU acceleration on your VDI deployment.
The addition of GPU cards does not necessarily reduce CPU utilization. Instead, it enhances the user experience and offloads specific operations best performed by the GPU.
Dell Technologies recommends using the BLAST protocol for vGPU-enabled desktops. NVIDIA GPUs are equipped with encoders that support BLAST.
Virtual Workstations are typically configured with at least 2 GB of video buffer.
Sizing considerations
This section provides general best practices for sizing your deployment. Not all of these options are configurable when ordering APEX Private Cloud and APEX Hybrid Cloud deployments.
- User density—If concurrency is a concern, calculate how many users will use the environment at peak utilization. For example, if only 80 percent are using the environment at a time, the environment must support only that number of users (plus a failure capacity).
Disaster recovery (DR)—For DR planning, Dell Technologies recommends implementing a dual/multi-site solution. The goal is to keep the environment online and, in case of an outage, to perform an environment recovery with minimum disruption to the business.
- Network isolation—When designing for larger-scale deployments, consider physically separating the management and VDI traffic from the vSAN traffic for traffic isolation and to improve network performance and scalability. This design illustrates a two-NIC configuration per appliance with all the traffic separated logically using VLAN.