Home > Workload Solutions > Virtual Desktop Infrastructure > White Papers > White Paper – Modernizing Retail with Scalable and Highly Available Virtual Desktop Environments > Virtual desktop load balance
With the demand for remote connection by users, the retailer described in this white paper decided to implement a highly scalable approach, leveraging service-load balance for user load balancing between sites. Dell Engineering deployed an NSX-T load balance solution to handle the load-based rules for each Cloud Pod. Each Cloud Pod works as a pool for users on each site, so we deployed a pod in each data center with the promise in the future to deploy more pods in logistic centers using VxRail clusters.
With this approach, the retailer has the capabilities of user-automated image deployment in the locations that are available. The Horizon infrastructure has the local load balancers for local management and a global load balancer that can handle the load between sites. Traditionally, disaster recovery and business continuity solutions have been used to protect mission-critical applications in the event of an outage, while virtual desktop solutions are designed for rapid image recreation.
With this architecture, we bring the optimal solution of combining robust disaster recovery and business continuity for virtual desktops with load balancers and Cloud Pod architectures. In an event of an outage, the user’s image is available in the second site almost instantaneously, with the VDI application components being handed to the second site by the VPLEX witness decision.
In the following figure, each Cloud Pod has dedicated infrastructure to run virtual desktops leveraging VxRail technology:
This approach with Cloud Pods and dedicated VxRail clusters brings greater flexibility for data center maintenance, networking topology, site infrastructure, installing new Horizon versions, or user load balancing between geographically dispersed sites.
The load balancer decides how the incoming traffic is distributed to all available sites with each Cloud Pod on it. The load balancer enables high availability for the virtual desktop solution and distributes the load among multiple sites in a totally transparent way to the end-users, while ensuring optimal resource allocation and maximizing throughput. The load distribution is achieved by mapping an external or public IP to a set of internal servers for data load balancing.
Several rules can be applied according to business needs, including: