The Unvarnished Truth About VDI? Not Hardly.

On June 14, InfoWorld published an article titled "The unvarnished truth about VDI desktop virtualization". In it, the author, Frank Ohlhorst, makes some claims about VDI and desktop virtualization that I believe need to be addressed.Ohlhorst starts off by saying that Virtual Desktop Infrastructure (VDI) will grant IT control over the desktop and succeed where thin-client computing failed. Not to nitpick, because I understand what the author is trying to say, but if IT does not have control of the desktop today, who does? Accounting? VDI will give IT better and more centralized control, especially if thin-client hardware is deployed. But if you've lost control of your desktops, you have bigger problems than VDI can solve.

My second issue is the author's assertion that thin-client computing failed. Again, Ohlhorst is addressing server-based computing, specifically Terminal Services and Citrix XenApp. Thin-client computing can be used outside of these technologies.

That being said, the fact is, there are over 80 million users of Terminal Services/XenApp today, worldwide. That's a failure? Ohlhorst seems to be saying that because Terminal Services and XenApp did not go totally mainstream, replacing every desktop, the model is a bomb. In fact, these technologies, while fully capable of replacing traditional desktops, were intended more to control the delivery of certain applications to users. In that respect, they have succeeded with flying colors. Major healthcare, financial and other sectors deliver selected applications such as Epic, SAP and Oracle ERP on a Terminal Services model with great success.

Most instances where these technologies did fail to meet the needs of the business were cases where the people responsible for the projects lacked the know-how to be successful or failed to plan properly. Neither Terminal Services nor VDI is a solution that you unbox, install, and figure out as you go. Unfortunately, many IT engineers ventured into such projects unprepared and hit brick walls. Any technology that has to do with the user community will typically be a high-profile application that generates a lot of noise and fury. Misconfigured environments lead to a user perception that the solution itself does not work, is slow. or not capable of the task. I believe Citrix is partly to blame, that the company should have been more insistent on a basic level of training for those embarking on XenApp deployments. But that was a business decision, one we expect Citrix may regret.

Desktop virtualization is shaping up to be déjà vu all over again. I see consultants every day architecting systems that are not based on carefully examined data. Poor planning is the main reason why desktop virtualization projects are failing; it's not that the technology is not mature enough.

The Challenges Of Managing Physical Desktops Today Is desktop virtualization really that complicated? Does it require that many more tools than physical desktop management?

Let's see... how many tools do you use today to manage your physical desktops? I am not sure they will fit in this blog if I start listing them. We can start with Microsoft System Center Configuration Manager and the list goes on, from security, backup, and antivirus all the way to imaging and troubleshooting tools and beyond. Managing physical desktops today is not a streamlined, automated, easy thing to do. The desktop is the most complicated computing environment you have in the organization.

Us server guys are smart, we silo applications on dedicated boxes or VMs and as a result, avoid application conflicts and ensure uptime However, the poor helpdesk team does not have that privilege. They have to deal with multiple applications, user data control, profile management, printing, and performance tuning and troubleshooting.

The answer, to my mind, is desktop virtualization. Note that I am using the term "desktop virtualization" because the InfoWorld article in question made a salad of technologies, for example, client hypervisors, blending them all under VDI. It is extremely important to understand that VDI is one form of desktop virtualization, but is not the only form. Client hypervisors and streaming are other forms of desktop virtualization. We can do one technology without doing the other. It is important that industry bloggers and authors do not contribute to the terminology mess that marketing gets us into sometimes.

Desktop Virtualization Is Ready Today The claim that desktop virtualization and VDI aren't fully baked just isn't true. While there are many organizations still investigating it, there are plenty of others that have embraced it and have figured out the most suitable solutions for their environments, from school districts and law firms to financials, call centers, healthcare and manufacturing.

Is today's desktop virtualization better than Terminal Services and XenApp? No, it's just different. Certainly the claim that was made in the InfoWorld article, that we could not put hundreds of users on a server and that it is more expensive, is flat out false. The user density that we can stack on a Terminal Services is still far greater than what we can do on VDI today. Furthermore, the InfoWorld author claims that replacing PCs with thin clients did not yield any cost savings. Well, of course not, if you decided to use the $550 thin client with an embedded Windows operating system. There are no cost savings there, but if you buy the $250 or less true thin client, you will see a significant cost reduction from a hardware, software, and management perspective. The TCO and ROI picture for Terminal Services and XenApp can be quite rosy, another area the author needs to brush up on. My take: Desktop virtualization is the natural evolution of Terminal Services and XenApp. It does solve certain challenges that we had in the Terminal Services days, such as that applications were not always developed for multiuser environments. Not every application was Terminal Services compatible. We could not assign resources on a per-user basis and so on, there is a large list.

But today's desktop virtualization has limitations and challenges compared with Terminal Services, too. Ohlhorst asserts that VDI has issues for which we are just seeing tools being developed. I will disagree here, as well. The challenges and troubleshooting scenarios that the author references are not new. They are the same issues that we had and still have with Terminal Services. After all, the remote protocols are the major bottleneck. While the author listed RDP and ICA, considering he mentioned VMware in many instances, he should have also referenced PCoIP. The author then talks about protocols and their consumption of bandwidth. Remote protocols can be very efficient and require little network bandwidth. That was the beauty of using them back in the day, because they worked over high-latency links. The question should be, how are you using these protocols? If you are using them for VDI on the corporate LAN, then you really don't care much, but if you are using them across the WAN or public internet, then you have to do some homework. What are the applications that the user will need? If it is just Office, then you don't need that much bandwidth. However, if the user will print, say, and printing is not configured properly, then yes, you will need a lot of bandwidth. The point I am trying to make is, no desktop is an out-of-the-box, turnkey solution. VDI requires proper planning and testing.

The question you should ask yourself is, are you happy with your existing desktop strategy? If the answer is yes, your costs are under control, backups and DR are not a challenge, moves and operational burdens are fine, then why change a model that is working for you? But if you are interested in exploring a desktop strategy that will reduce costs, enhance IT's grip on desktops and offer better ways of addressing old issues, then you should absolutely take a look at desktop virtualization.

Desktop virtualization is here today, there is no need to wait. IT is up against a Windows 7 upgrade and hardware refresh on the desktop. The choices are simple: Invest in new fat hardware and management software and continue to do it the old fashioned way, or leverage virtualization and explore new, better, and faster ways of managing desktops.

Elias Khnaser is the practice manager for virtualization and cloud computing at Artemis Technology, a vendor-neutral integrator focused on aligning business and IT.