Many companies are turning to Virtual Desktop Infrastructure (VDI) as a solution for providing their global workforce with easy access to work environments. Using VDI technology, companies can consolidate desktop operating systems, applications, and data into a centralized data center, which employees can access from any location using virtual desktops.
Scaling VDI is fairly easy, but the scalability of the hosting IT infrastructure is another matter. As your company adds more users and virtual desktops, you must also add more compute, storage, and networking power to support the high-density, IO-intensive workloads of VDI and maintain the proper user experience in the process. IT scaling is often a complicated and expensive task, and it can leave you with a complex IT footprint that reduces the ROI of your VDI system.
In this article, we’ll explore how having the right infrastructure can help you to overcome the major problems of scaling your infrastructure to support your VDI initiative, both now and as your needs grow.
The traditional three-tier architecture used for hosting VDI includes separate servers, storage, and storage networking, creating separate points of management and silos in your datacenter. As you scale up, you must add more resources to each silo. Over time, this makes your IT footprint for VDI very complex and harder to manage.
But Hyperconverged Infrastructure (HCI) eliminates these silos by combining compute, storage, storage network, and virtualization into modular, software-based appliances, or nodes as they are referred to in HCI terminology. This allows you to scale in a modular, building block fashion that simplifies scaling.
You should look for an HCI solution that gives you the ability to manage your entire IT footprint for VDI from a single user interface. This makes it easy to scale out in multiple pre-defined pods of VDI users, yet still have a single management point, thus ensuring simple IT management as user count continues to scale.
Higher Costs of Scaling
Scaling traditional IT infrastructure also drives up costs, since you must buy additional separate types of hardware and software each time you expand your VDI deployment. Since HCI is based on less expensive industry-standard servers, it cuts down the cost of IT scaling.
Also, look for an HCI system that utilizes an ultra-low latency NVMe flash storage tier to drive high-density deployments and optimize the user experience. This provides even further cost savings because you can host up to 2-3 times more virtual desktops on each node. In other words, you can host more desktops while buying less hardware.
Unpredictable Scaling Needs
As your company’s workforce grows, it’s hard to predict the compute and storage capacity your IT infrastructure will need to support future growth on your VDI system. Many companies end up over-provisioning resources, making a large investment in new hardware in order to accommodate future VDI user growth.
Look for an HCI solution that allows you to deploy and manage multiple “pods”. A pod can designate a certain number of VDI users and nodes per pod (i.e. 500 VDI users = 1 pod = 4 nodes), which makes it easier to predict scalability needs. Instead of adding nodes one by one, you can add pods and be confident of the number of VDI users (i.e. 500) that pod will support successfully.
Inflexible Scaling Options
A common problem with many HCI systems is that compute and storage resources are often tied together in the purchase of an HCI node, and you can’t buy one without the other. After scaling up their HCI platforms, many companies find themselves with a surplus of compute or storage for VDI.
Therefore, it makes the most sense to look for an HCI solution that provides flexible deployment options, allowing you to scale up in granular dimensions, according to your needs. You should be able to add the right combination of individual compute, storage, and/or GPU nodes to your IT infrastructure, in order to address user needs, prevent bottlenecks in your VDI system, and scale in the most cost-effective way possible.
VDI Performance Issues
Scaling itself can be a disruptive activity. In traditional IT infrastructure, upgrading your SAN system usually requires you to take the VDI platform offline, which interrupts VDI service to your employees. That’s why it’s critical to select an HCI solution that allows you to scale up in a non-disruptive manner.
Also, many HCI systems operate on a principle called “data locality,” which limits the system resources available to virtual desktops to what is contained in the HCI node where the virtual desktop resides. This can create performance issues such as latency, which will impact the user experience.
Look for an HCI solution that pools resources from all the nodes and makes them available to all virtual desktops in that cluster. Also, the HCI solution should automatically re-distribute volumes and data sets across all nodes in a cluster each time you add a new node. This re-balancing of data increases IO capacity and helps to prevent performance issues as the IT infrastructure scales to accommodate more VDI users.
As your VDI system scales up from hundreds to thousands or tens of thousands of users, it’s important to ensure that the hosting IT infrastructure is robust and flexible enough to support your VDI workloads. But scaling your IT footprint doesn’t have to be an expensive or disruptive experience. If you use an HCI infrastructure that provides a distributed, scale-out architecture with flexible, granular deployment options and architecture optimized for performance with NVMe flashit becomes easier to plan for scaling needs, simplify your IT footprint for VDI, and keep costs under control.