Some technologies go through a hype cycle that results in the positive aspects of that technology being viewed in an imbalanced way when compared to the potentially negative aspects of the technology. This results in a skewed favorable look at the technology, which serves to mask its potential pitfalls. Virtual Desktop Infrastructure (VDI) is one such technology. Here, you will discover four items to take into consideration before you take the VDI plunge.
You will not save money with VDI
This is a statement made with certainty: In most cases, if you are undertaking a VDI initiative for the sole purpose of reducing expenses, stop your efforts. VDI does not save money. If it does save money, it’s done at the expense of the user and it’s more than likely that the effort will be perceived as a failure.
When VDI is done right, the user experience will not suffer. Doing VDI right means investing money in core infrastructure, which includes robust VDI-tuned storage, servers and good terminal devices. Between these infrastructure needs and the VDI licensing that is incurred, organizations will generally and at best break even when comparing a new VDI environment to a traditional well-managed desktop environment.
For those that are under particular strain to reduce ongoing costs for the desktop computing environment, there is a shortcut that might be taken, but it comes with some tradeoffs: It’s entirely possible to use existing desktop hardware as terminals. However, this will negate any energy savings that might be wrought from a VDI initiative. Further, since Windows will still need to be installed locally, IT must maintain two Windows images—one local and one virtual—for users, adding significant complexity to the environment.
Finally, a desktop PC has a lot of moving parts that can still fail. If you’re planning to do VDI, ensure that you’re doing so for value-add business reasons that don’t involve saving money. Perhaps you’re deploying VDI to support a BYOD initiative or to provide secure desktops to a mobile sales force. Or, perhaps you’re a college or university and want to create an on-demand lab for your distance learning student population. Regardless, there must be a business-facing reason for deploying VDI.
Storage is the Achilles Heel in VDI deployments
In a traditional server virtualization scenario, storage plays an important role, but RAM is generally the resource of most concern as the first resource to be exhausted. In VDI scenarios, RAM is obviously important, but storage is the first resource to go, but not in the way that you may think. Due to various means available to control runaway storage needs (such as linked clones), capacity isn’t always a problem. Instead, storage problems usually revolve around storage performance issues that are extremely visible to the end user.
Many VDI storage performance calculators do an exceptional job at predicting normal performance needs as a user is performing his job. Unfortunately, though, the Windows operating system isn’t the most efficient software when a system is booting and the user logging in. During these times, VDI environments may suffer from what are known as “boot storms” and “login storms” during which high numbers of users are attempting to boot or login at the same time. In the worst case, this results in it taking tend of minutes or even hours for virtual desktops to make it to a ready state.
Adding to the challenge is the fact that the Windows boot and login process is write-intensive. Due to RAID striping and parity overhead, writes carry significant performance penalties that must be considered.
If you’ve discovered a compelling business need for VDI, make sure that you architect your storage environment to withstand these storms. This often means combining traditional storage with high-IOPS solid-state disks (SSDs). SSDs carry IOPS values that range into the thousands per drive, making these expensive devices well-suited to helping alleviate boot and login storms.
Another option is to consider some of the purpose-built VDI solutions that take a building block approach and combine both traditional and SSD drives in each building block. These solutions carry two benefits:
- They address the massive upfront costs that can be required in VDI deployments by allowing organizations to spread the cost out over the full desktop replacement cycle rather than having to buy the core all at once. This can be a huge benefit when it comes to adhering to lifecycle plans in the organization.
- These solutions are purpose built to support VDI. That’s what they do.
There are new licensing challenges to consider
If anything will completely derail a VDI initiative, it’s software vendors using VDI as a way to improve the bottom line. Even if you already have site licenses or concurrent user licenses for your software titles, if you plan to deploy VDI, you need to review all of your license agreements to determine whether or not they include a provision for virtual execution.
Many software licenses specifically prohibit use in a virtual environment and the vendor requires that you purchase specific virtualization rights in order to use the software in VDI.
In addition, pay special attention to Microsoft licensing with regard to VDI. The company sells a license known as the Virtual Desktop Access (VDA) license that is to be used to conjunction with virtual desktops accessed from non-Windows-based devices, such as thin clients and tablets. This is a per-device license, which is a major disappointment, especially in BYOD scenarios or when users want to access their virtual desktops from many devices.
Microsoft makes available a comprehensive Product Use Rights document that outlines what organizations can and cannot do with the company’s software. Review it in depth and then call your Microsoft software reseller and look for an expert. This is an extremely complex area.
For other titles, I wish there was specific guidance I could provide, but your best bet is to get your company’s best negotiator to work out details on a per-product basis.
Is your deployment about the desktop or is it about the apps?
I’ve seen organizations considering VDI deployments for mobile workers in order to provide them access to a core application or set of applications. Before to pull the trigger on your VDI deployment, ask yourself an important question: “What problem am I trying to solve?”
If the answer is nothing more than remote access to applications, VDI is the wrong solution. In this case, the stalwart Terminal Services/Remote Desktop Services that has graced the IT landscape for years is a much better solution for a number of reasons:
- It’s easier to deploy.
- It’s a well-known use case for software licensing.
- For applications only, it will scale much better (no multiple OS instances, for example)
The most important VDI question to answer is whether you need VDI at all. If your goal is to provide access to applications, there are far simpler ways to get there.
Action Item: VDI has enjoyed a lot of hype and, when done for the right reasons and built the right way, it can transform a business, expand the reach of a college and make things more accessible for all. However, there are pitfalls that can’t be ignored. Before you leap into the VDI ocean, look for any sharks that might be circling.
- Do VDI for the right reasons.
- Build a storage environment that addresses VDI challenges.
- Consider any and all licensing pitfalls.