Overview
VDI solutions based on Dell PowerEdge, Dell storage, and Dell PowerSwitch networking provide flexibility as you scale, reducing the initial and future cost of ownership. Add physical and virtual servers to the server pools to scale horizontally (scaling out). Add additional resources to the individual components of the infrastructure to scale vertically (scaling up).
Scaling out
Each component of the solution architecture scales independently, depending on the required number of supported users. You can add servers, storage, or networking at any time to expand the architecture in a modular fashion. The scaling limit for a vSphere cluster is 96 nodes per cluster.
Sizing recommendations change over time as updates are released and qualifications are performed. See the VMware Configuration Maximums website for the latest recommendations.
A Citrix "pod architecture" consists of a virtual desktop "site" and one or more "zones" at each site. The maximum limit when using Citrix pod architecture with the LHC feature enabled would be 10,000 VDAs per zone and 40,000 VDAs per site.
A reasonable limit to choose is 5,000 linked-clone VMs per zone. With this limit in mind, 25 compute nodes with 200 task-user VMs per node would reach the maximum number of VMs for the zone. This Dell Validated Design for VDI uses MCS linked-clones. The following figure shows a 5,000-user MCS linked-clone configuration, which equates to a single zone within a site:
The following figure shows a scale-out of 20,000 MCS linked-clone VMs. The recommended limit for a zone is 10,000 sessions. In this example, there are two sites with two zones each. The sites are load-balanced and presented to the users as a single entity.
Scaling up
Each of the components within this design can be scaled up in the following ways:
- Dell PowerEdge servers
- Processors—If more processing power is required, higher core count and clock speed processors can be installed up to the highest model.
- Memory—Additional memory can be installed into each compute node up to the limit of the system architecture.
- GPUs—Additional or more powerful graphics cards can be installed up to the capacity of each compute server.
- Dell storage
- Storage Controllers—Many of the Dell storage systems support scaling up the storage controllers by swapping, clustering, and generational upgrades.
- Storage capacity and performance can be improved by adding additional disks to storage systems in empty drive slots or by adding expansion shelves.
For more information about Citrix best practices for scaling, see the relevant content in the Citrix Tech Zone.