Virtual Desktop Infrastructure (VDI) is rapidly becoming the solution of choice for organizations looking to simplify and streamline their End-User Computing (EUC) operations. By virtualizing and consolidating end-user operating environments, VDI simplifies the lifecycle management of user desktops while enhancing security and boosting employee productivity. However, the long-term success of VDI deployments is critically dependent in choosing right infrastructure for your VDI.
In this blog post, I will discuss some of the risks the underlying infrastructure can pose to VDI deployments and the best approaches to mitigating those risks. However, to do so, we need to understand the fundamental nature of VDI workload from infrastructure perspective.
VDI is one of the densest workload IT organizations support. In VDI environments, a large number of desktop virtual machines are virtualized on a few server nodes and they share underlying storage access. This is one of the most IO dense and demanding environment for underlying shared storage, with extreme burst conditions and IO storms a commonplace. VDI projects are also known to be implemented in phases, meaning VDI scales over time to accommodate larger user pools. Most of the risks to your VDI can be attributed to the extreme nature of VDI workload.
Risk #1– Inconsistent and Unacceptable User Experience: User Productivity and VDI Adoption at Risk
VDI adoption, and as a consequence, the long-term success of VDI projects, is dependent on ensuring that end-users who use those desktops are happy and productive. User experience is closely correlated to the IO performance and latency of underlying storage resources. In events where a large number of desktops access storage simultaneously, such as during morning login hours or during the busiest hours of the day, many traditional infrastructures are unable to service those requests in a timely manner and thus cause the user experience to be sub-par, and even severely affect their productivity.
Having a sub-par and unpredictable user experience is a non-starter. And it’s the infrastructure that really dictates the quality and consistency of user experience. So, watch out for this. Look for infrastructure platforms and technologies with built-in intelligence that help you prioritize user pools and datasets and enforce performance SLAs. Look for architectures designed around faster storage technologies like NVMe flash that can help overcome IO bottlenecks and thus provide a better user experience.
Risk #2– Cost and Complexity of VDI: ROI at Risk
Unlike virtualizing servers, which usually defers or minimizes new hardware purchases by consolidating more applications on a fewer nodes, VDI entails deploying new and high-performance datacenter hardware to support all those desktops. Thus, managing cost, especially related to hardware footprint, is a key imperative to meet ROI objectives. However, due to the high-density nature of the workload, infrastructure and storage in particular is often overprovisioned to ensure system performance and user experience. To make matters worse, newer operating systems like Windows 10 require significantly more resources and thus more hardware to support the same number of users. By some estimates, Windows 10 can reduce per node desktop density by up to 50%, meaning it can require up to twice as much hardware. This can compromise cost structure and ROI.
Keep a close eye on the infrastructure footprint and look out for infrastructure overprovisioning in your VDI. Look for infrastructure platforms that integrate newer faster storage tiers in their storage architectures that can help deliver better user densities per platform. In general, look at hyperconverged infrastructure (HCI) since HCI eliminates the need for specialized SAN or NAS storage systems. It uses industry standard hardware, and it can significantly lower the cost and complexity in your VDI. Evaluate the HCI choices by how many users they support per node/system, in particular your user type.
Risk #3– Unpredictable, Complex Scaling: Growth Planning at Risk
VDI deployments usually scale in multiple phases. Initial deployment may be a pilot supporting a small user pool, that can scale through regions and functional areas in many phases as VDI is tolled out throughout the organizations. It is common to see VDI start with as few as 100 users and scale to many 1000s of users over the next few years. The challenge, however, can be with scaling underlying infrastructure to support this growth. Often, scaling can be disruptive, causing service outages. Many times, scaling to accommodate small user growth can require large up-front investment in infrastructure. System performance and use experience can become issue as the scale grows, and thus the operational overhead of managing performance at large scales. The infrastructure itself can become too complex to manage at a certain scale, adding administrative overhead.
All of this can stymie broad roll out of VDI projects. You should look for a platform that scales predictably, non-disruptively, and in just the dimension needed (storage, compute, GPUs, etc.). Some platforms may offer you the ability to scale in small, granular steps and address specific resource bottleneck as you scale. This could significantly simplify scaling exercises by offering flexibility and lowering cost. Pay close attention to manageability of the infrastructure at large scale. Platforms offering POD like manageability and scalability can help keep it simple at scale.
VDI is transforming end-user computing, with benefits that help simplify management, improve security and enhance user productivity. However, close attention to infrastructure is necessary. Inadequate infrastructure can introduce risks related to your ROI, user adoption and broad roll out of VDI. New infrastructure technologies, particularly HCI platforms with modern architectures that utilize faster storage tiers, provide SLA management tools and provide flexible predictable scalability, can help address these risks and help IT deliver a successful VDI.