If you find yourself in a similar position, here are a few good high level technical questions to ask yourself as you research and prepare for a VDI project:
Is data security a key concern for your business?
While the immediate answer that comes to mind is "yes", stop and think about it before answering. If it is important to add another layer of protection for your data, ensuring it never leaves the data center even with users scattered around the planet, that might make VDI worth it all by itself.
What types of users do you have?
Are they all in one spot? Are they all over the world? Does it make sense for some to be able to work from home as well as the office?
What applications will get virtualized?
Most applications are perfect candidates for desktop virtualization (VDI) but some are painful or flat out impossible to work with. This would be a good thing to know before you get neck deep in a VDI project.
What desktop OS are you going to deploy?
Microsoft could make all of this very easy but they choose not to. As a result we have some extra hoops to jump through to stay in compliance with Microsoft licensing and still deliver a cost effective solution. Odds are you are going to get a server OS that looks a lot like Windows 7 or Windows 10, but will not be the real deal.
Still have legacy apps requiring Windows XP? That needs to be thought through, but is not a show stopper.
Are your applications 64 bit aware, or legacy 32 bit?
The farther we get from Windows XP and Windows 7, the less of an issue this is, but just in case it is worth mentioning.
Will your end users need remote access?
Having your desktop available to you on any device, anywhere at any time is often one of the key considerations for virtual desktops (VDI). Having your desktop seamlessly follow you on your tablet, smart phone, desktop, Macbook or Chromebook no matter where you log in from can be a powerful reason to look at desktop virtualization.
What are the storage implications of virtualizing your desktops? Do you have the space?
Storage space, speed and IOPs can be a key consideration. SSD drive technology has fundamentally changed the data storage business but this is not something to be overlooked.
3 Big Challenges to Desktop Virtualization
1. Software Licensing Terms can have a massive impact on the viability of a VDI project. Microsoft licensing is one key area that must be evaluated. For VDI deployments, Microsoft has issued special licensing guidelines. For VDI you will need a Virtual Desktop Access (VDA) license. A normal for-profit enterprise can expect to pay anywhere between $100 a year and $360+/- for a multi-year agreement per covered edge device. If your organization is under a Microsoft Enterprise License Agreement (ELA) or you have Software Assurance, the cost of the VDA license is included under your licensing agreement, thus your additional cost per end device to zero for VDI. Note this license is in addition to each user needing a Windows Server Client Access License (CAL) and a Terminal Services or now Remote Desktop Services CAL if you are deploying your VDI desktops with Citrix or Terminal Services-esque technology. Want a deeper dive into Microsoft VDA licensing, click here.
If VMware is your hypervisor of choice, you are likely to face some significant software cost increases here as well.
2. Uncertain ROI More often than not a VDI project will not bring a significant CapEx savings over buying PCs. You may save a little of thin clients are in your future or if you are able to extend the life of existing hardware. However, new servers, potential network upgrades, new desktop devices (i.e. thin clients,) and data center storage might also be in your future and will need to be factored into your VDI ROI model. The real ROI, if it is there to be had for you, will likely come in the form of more OpEx costs like:
Keep in mind that VDI deployments will almost always be more expensive than publishing applications via Terminal Services or Citrix. VDI-in-a-Box was one possible exception to this rule, but Citrix has discontinued this product.
3. End user Acceptance The VDI end user experience for end users needs to be AT LEAST as good as their old trusty desktop to get end user acceptance of the project. This is a critical component that directly impacts the success and viability of any VDI project and I find it is the least measured or accounted for. There will be no ROI if you end up implementing a d esktop replacement technology that does an exceptional job or reducing employee productivity or if the end users revolt.
The end user experience needs to be proactively managed and monitored to make sure the projected benefits of VDI result in happy employees.
We like to see login times to be 30 seconds at a maximum in VDI environments, with 15-20 seconds considered average, and 5 seconds +/- would be considered flaming fast.
1. Get a VDI Readiness Assessment A good VDI Readiness Assessment will crawl your network in advance of your VDI project and deliver a report on where your problem areas are before you start. This will also help you find and understand any licensing restraints.
2. Deploy VDI only where the use case makes sense The best way we have found to achieve ROI is to blend VDI into your environment where it makes sense and use other methods for deploying applications where VDI does not make sense. For users that need to be able to install software, have unique application requirements, etc., look to VDI, for the balance of the user population, look to XenApp or Terminal Services to give you better user densities and cost savings.
3. Provision Desktops Leverage provisioning technology and create/maintain a few desktop images for your users instead of having to maintain each and every one. Managing a handful of server and desktop images instead of hundreds of PC images can be a very easy OpEx conversation.
4. Deploy zero/thin clients Convert existing PC’s to stripped down thin clients or deploy zero client or thin client technology with embedded management tools to improve access and remove the need to visit the edge device.
5. Review and Improve the End User Experience From my experience, this is the most overlooked aspect of planning and delivering a VDI deployment. The end user experience should improve with VDI, not get worse. A poor end user experience leads directly to poor adoption rates and increased end user resistance.
We want to improve productivity and reduce costs. Having no visibility into the end user experience will not allow you to consistently achieve your productivity and savings goal. We use some proprietary tools, as well as ControlUp and Aternity to provide us visibility from any edge device (iPhone, iPad, Android device, Mac, laptop, etc.,) including key metrics for each application, back through every component in the data center. This gives us a complete proactive view of our end users so we can see issues before they pick up the phone and call support.
6. Implement High Availability for Desktops Build redundancy into your VDI model so that end users cannot lose their desktop. PC’s crash, blue screen, grind hard drives to a halt and smoke power supplies. VDI gives you an opportunity to improve the end user experience by giving them a desktop that will not die. Well, will at least be more difficult to kill.
According to Aberdeen Group research on desktop virtualization, 79% of Best in Class performers monitor desktop application performance, while 50% monitor end user satisfaction with their desktop devices and 69% have deployed thin clients in their environment1. Total desktop downtime was reduced by 36% and time spent on routine desktop maintenance was reduced by 28%1.
(1) Aberdeen Group (2012) “Best-in-Class Companies Choose Thin Client Desktops”