What is Virtual Desktop Infrastructure?

Not many years ago, computing at work was simple and inflexible. People walked into their offices, punched the power button on their desktop, tapped their fingers on the desk waiting for it to start, and then got on with the applications, files, and web browsers as and when needed. They weren’t much comfortable with the process but eventually got familiar with the system.

For the enterprise, though, this was a great headache. In addition to paying for hundreds or even thousands of desktop or laptop workstations (with all the software required) for each employee, the system admins also spent considerable time doing hardware maintenance and updating software.

Not just large companies, but hospitals, schools and government agencies also faced the same problem – scaling up computing, storage, and networking capabilities also resulted in a proportional increase in costs.

Virtual Desktop Interface (VDI) technology came to the rescue around 15 years ago. Today, it is part of the larger End User Computing (EUC) ecosystem and involves a centralized OS that is hosted, run, and managed in an on-premises data center or in the cloud.

Virtualized environments called Virtual Machines (VMs) are then deployed over a network to end users who access them using a variety of endpoint devices. These VMs contain a set of software and hardware resources defined by the users’ roles, and can be accessed as if they were running locally on the users’ devices.

How Does VDI Work?

Technically speaking, virtualization is a software-based, virtual representation of a collection of servers, storage, applications, and networks, where the underlying physical devices are abstracted by software called a hypervisor that runs on top of the operating system and is managed via a single user interface.

This hypervisor decouples the underlying hardware from the logical OS, which runs on a server in the data center and facilitates multiple virtual desktops to be served from that server. These virtual desktops are served to individual user instances via a software platform called a connection broker, which serves as a gateway between the hypervisor and the user and their endpoint device.

There are two central characteristics of every VDI implementation:

• The virtual desktops are “host-based,” which means they are all housed in a central server in the data center. Each VM contains an OS image (Microsoft Windows is more common compared to Linux) and is allocated access to limited hardware resources connected to the server.
• An unbroken and network connection between the server and the end client is needed to ensure continuous access to the virtual desktop. The connection broker assigns an appropriate virtual desktop from the central resource pool to each client, while the hypervisor creates and manages each VM that encapsulates the individual VDI environment.

It’s crucial to note that VDI is different from desktop virtualization, which is just a single desktop device running a single guest VM instance. VDI bypasses the OS of the local computer and delivers a shared desktop interface instead.

Different Approaches to Deploying VDIs

Virtual desktops are served in two flavors: Persistent and non-persistent.

Within a persistent VDI, users are allocated the same desktop images every time, enabling changes they make to apps and files to be retained in the VM. The first time a user logs in, they’re assigned a standard desktop or “golden image” by the hypervisor. On each subsequent login, they’ll be connected to the same desktop, allowing for a personalized environment – the changes they make to various apps and settings (such as passwords, shortcuts, and files) are all retained.

Thus, persistent VDI ensures the same level of flexibility and familiarity to users as their own desktop or mobile devices. However, there are management and hardware costs to persistent VDI. A 1:1 desktop to user ratio means there are as many separate images as users for IT to manage. The storage cost for each user and image rises accordingly.

In contrast, a non-persistent VDI serves up fresh desktops to all users upon login and doesn’t allow them to save any changes. While this ensures a minimal amount of images for admins to manage and saves storage costs, it limits or constrains every user’s flexibility on an individual level.

Advantages and Disadvantages of a VDI Deployment

First, the benefits:

• VDI saves a lot of system administration time. IT staff don’t need to spend hours installing OSes or apps, or patching and updating software on end users’ machines. Ditto for backup. All data and applications are stored and accessed from a central location, simplifying backup and restore processes.
• There is little need for computing or storage, if any, at the endpoints. This lets IT repurpose old or less powerful desktops as endpoints. When there is a need for scaling up the number of users, organizations can purchase and allocate thin clients to them or even let them use their own devices, given the recent surge in BYOD and remote work practices.

This is especially useful when a company wants to let agents, contractors, or temporary workers into its network and have them access a particular workload or application.

Case in point, when the coronavirus turned into a pandemic, the staff at Hastings Prince Edward Public Health (HPEPH), a public agency offering programs and services to citizens across Ontario, Canada, had to switch to work from home overnight, per government order.

“We went from business as usual to having to ramp up a large call center in two days,” said Tom Lockhart, IT Systems Manager at HPEPH. Fortunately for him, the infrastructure for a VDI on Nutanix was already in place, prompting him to claim, “If we need 100 new sessions, I just provision them and they’re ready to go right away.”

• Disruption caused by endpoint failure is greatly reduced. In the event of a problem with the endpoint device, the user can quickly switch over to another readily available device without worrying about minimum requirements.
• VDI offers significant benefits in security, given that critical data is always stored on the server. Unless an attacker has infiltrated a live session, there is no danger of data loss in the event of device theft.
• The biggest advantage of VDI is that it allows for a consistent and flexible user experience – regardless of whether they’re using a smartphone, laptop, or kiosk to access the workload – while increasing their mobility.

VDI also has its drawbacks, but they are mostly specific to the technology:

• Here, user experience comes first. Some users might get confused between their own local desktop and the virtual session, rendering them susceptible to making trivial errors such as attempting to save files in the wrong location.
• The compute, storage, and software licensing costs of the server tend to rise quickly as the number of users and their application-related requirements increase. Workload capacity and bandwidth needs will also balloon up proportionally.
• VDI infrastructure is very complex and warrants a lot of hardware and software components working in tandem. The centralized server is a single point of failure. If the hypervisor, broker or licensing server fail, or the server runs out of storage space, the entire system goes down and users are locked out of the workload.
• A lot depends on the internet connectivity speed of the end user. While this isn’t as big a problem as it used to be, slower connections can quickly frustrate users, lead to delays and lost productivity, and affect mission-critical application and database transactions.

VDI Use Cases

A large number of organizations cutting across industries have deployed VDI – and even ramped up its scale – in the aftermath of the COVID-19 disruption.

Here are a few work environments where VDI deployments can definitely add efficiency, speed, and productivity:

• Schools and universities: VDI in schools means more opportunities for delivering and receiving education – classrooms can be streamed to students’ homes to enable continued education, especially in the post-pandemic era. Students and faculties can be issued devices for use during their tenure at the institution.

For instance, California’s Oakland Unified School District distributed Google Chromebooks to 25,000 low-income students. Since each Chromebook costs hundreds of dollars less than a traditional computer, school districts can bridge the digital divide without exhausting their budgets. Knowing that some students don’t have reliable internet access, Oakland Unified also distributed portable Wi-Fi hotspots. 

• Call centers and other companies that work in shifts: In an organization with a task-based workforce that flexes up or down, shared desktops are the norm. Employees can simply log on to a simple, non-persistent VDI, bring up a standard desktop with a limited set of apps assigned to them, complete their tasks, and log off at the end of their shift.
• Highly regulated environments: In industries such as legal, healthcare, finance, military, and telecommunications, VDI provides the assurance of security as well as granular control over each user’s desktop. This eliminates the possibility of end users storing sensitive data on a private machine or that of their devices being hacked.

“It took me two to three years to get through most of the red tape. Once I was able to quantify the risk regarding $28 billion in assets and provide clear and concise information regarding risk, we finally got the buy-in that was needed. We simply could not put the complete set of security controls in place with the way it was architected. Since then, from an on-prem, hybrid infrastructure perspective, we have settled on Nutanix, at the core,” said Michael Wyatt, lead architect at Vodafone, after implementing a hybrid infrastructure involving VDI, HCI, and the public cloud, with a ZTA security plan for over 50,000 users.

• Engineering and design companies: Tech companies – even startups – are known for their use of compute-heavy applications in CAD/CAM, graphics, or software development. These are also environments where BYOD is most prevalent. Almost every user is a power user and needs a non-standard configuration. Recent developments in VDI technology and reductions in hardware costs have made such resource-intensive work possible on VMs.

Regardless of the nature of deployment or use case, the most important factor to the overall success of the VDI deployment is clear and open lines of communication with end-user stakeholders, according to Yangzhi Zhao, Product Director of Nutanix Frame. 

“First and foremost, IT should work closely with all stakeholders to ensure they have a solid understanding of end user requirements and conduct proper segmentation of users into different use case groups. Communication also includes educating end users on the benefits, differences, and limitations of virtual apps and desktops solutions early on (and throughout) the deployment,” he said.

DaaS: VDI in the Cloud

Over the last ten years or so, an “as-a-Service” component called DaaS has been added to the VDI kitty, proving to be a savior for SMBs that lack the resources to build and deploy a VDI.

The critical difference between VDI and Desktop as a Service (DaaS): While VDI is typically implemented on on-premises data centers, DaaS enables End User Computing (EUC) from a public or private cloud infrastructure, requiring less customized or specialized configurations. DaaS also lets employees run essential apps from their own devices while freeing admins up from monitoring every session and updating endpoints.

Here’s a simple and fun animation that captures the differences between VDI and DaaS in a story-like form:

The choice between VDI and DaaS starts with a crucial question: How do organizations choose between cloud or on-premises deployment based on cost/complexity considerations? There are a lot of overarching variables and factors – such as use case, data locality, user locality, security, licensing, and supportability – that IT needs to take into consideration when it comes to deciding whether to deploy their virtual application and desktop workloads on-premises, in the cloud, or both (in a hybrid model).

Zhao said it’s important to focus on the use case.

“How will your end-users be consuming virtual apps and/or desktops?,” he said, suggesting how to access the situation. “Is the use case a desktop replacement scenario for 9-5 task workers? Is the use case for a knowledge-intensive group of power users for a short-term project?” 

Generally speaking, for use cases that require more consistent and dedicated usage patterns, deploying those workload on premises in dedicated hardware makes sense, he said. 

“On the other hand, for more dynamic use cases where usage patterns can vary significantly from month-to-month, deploying to public cloud may be a better fit.”

Elaborating on the cost/complexity dilemma, Kong Yang, Senior Marketing Manager for EUC at Nutanix shed some light.

“What businesses really want to understand when looking at OPEX vs. CAPEX for VDI and DaaS, is how do we meet our company’s financial goals while delivering excellent user experience at scale?” said Yang.

“From an OPEX perspective, customers who go with DaaS don’t want to be surprised by a bill that’s much larger than expected. On the CAPEX side, they don’t want to be short-changed by the solution and forced to buy more infrastructure to meet scalability requirements,” he added.

So is DaaS set to replace VDI eventually? Not really, believes Nikola Bozinovic, general manager of desktop services at Nutanix. 

“VDI and desktop as a service are two things that coexist and will continue to coexist for a long time,” he said in an interview with The Forecast. “VDI assumes that everything is contained and delivered as software, whereas DaaS is hosted by a vendor, so the service is always up to date and secure.”

Hyperconvergence: The Future of VDI

Given that the world’s biggest work-from-home experiment is currently underway, nearly every organization, small or large, is using a hybrid cloud infrastructure in one way or another.

“Just like the future of cloud, the future of VDI is hybrid and there will continue to be different deployment models to address different use cases. There will be customers, for security reasons, who will need to remain on-premises in a private datacenter, and traditional VDI will continue to be the solution for them,” said Zhao.

“For most customers, however, who will operate in both private and public clouds, DaaS solutions that support hybrid and multi-clouds will be the answer,” Zhao said.

That said, VDI’s promise of a simple and cost-effective unification of hybrid cloud systems and on-premises data centers is now being fulfilled by hyperconvergence. It delivers compute, storage, and network solutions transparently to the end user while keeping complexity to a bare minimum. This means IT teams can spend less time worrying about the infrastructure and focus on how the workloads running on top of it can add value to the organization.

Dipti Parmar is a marketing consultant and contributing writer to Nutanix. She writes columns on major tech and business publications such as IDG’s CIO.com, Adobe’s CMO.com, Entrepreneur Mag, and Inc. Follow her on Twitter @dipTparmar and connect with her on LinkedIn.

© 2021 Nutanix, Inc. All rights reserved. For additional legal information, please go here.