Brian Duckering – Marketing Strategist

PART 3:  Endpoint Virtualization Series

Maybe you are determined to buy some virtualization technologies, but are having trouble figuring out where to start – or where to end. The first question to ask is ‘Why am I doing this?’ What do I hope to accomplish by virtualizing? Am I sacrificing anything that is important to my business in the process? Then you can determine the best way to accomplish that, realizing that the solution may come from the world of virtualization, or not. I will again stress the importance of reviewing each type of virtualization independently of the others, as they solve very different problems and have unique cost models and value equations.

Let’s start with a quick (grossly over-simplified but useful) summary of value for the three levels of virtualization previously discussed.

• Server Virtualization: Reduce hardware cost and increase flexibility in the datacenter.
• Desktop Virtualization: Reduce cost of deploying and maintaining end-user systems.
• Endpoint Virtualization: Automate end-user workspaces and reduce IT support costs.

Note that cost reduction comes up a lot when talking about virtualization. This is generally a reasonable assumption, but note where additional training and skills will be necessary to support the new technologies. Plus, virtualization rarely completely displaces pre-existing or non-virtual technologies, so it may also be an additional cost, even if that portion of the budget is smaller. I will not spend a lot of time on Server Virtualization, first because it is generally better understood than the others, and second because most of the discussion and confusion today is around the other two. Here is an easy distinction: THE DESKTOP IS THE WORK ENVIRONMENT, NOT THE WORK ITSELF. While Desktop Virtualization provides a different door to a different office – the user will still need the same things he always needs to do his job, whatever office he may be using today. Endpoint Virtualization is about the THINGS the user needs to have IN the office to get his job done, also referred to as his WORKSPACE. Today, I’ll just focus on Desktop Virtualization.

A brief overview of Desktop Virtualization

As mentioned above, Desktop Virtualization is about reducing the cost of deploying and maintaining end-user systems. Traditionally, client systems are physical laptops and desktops, and much of the cost comes from the variety of hardware profiles (increasing support costs and reducing stability) and the distributed nature of the systems (making IT access more difficult and costly). Virtualizing these desktops simply relocates the client operating system back to the datacenter, improving both complicating factors. Now the client desktops can be in one easy-to-access location on similar or identical hardware. [Remember that even when the client operating system is in the data center, some form of distributed hardware is still required for user access.] There are three common approaches to this, each with its own cost model. They are, in order of increasing overall cost: terminal servers, virtual hosted desktops (or VDI), and blades. So why doesn’t everybody just implement terminal servers for everybody since it is the cheapest? The answer is that there are tradeoffs and compromises with each solution.

Choosing the right client desktop model is a matter of evaluating and balancing needs – both the needs of IT to be manageable, cost-effective and secure; and the needs of the end user to be productive, connected, mobile, flexible, personalized, etc. Evaluating different groups or departments within a company should yield different conclusions. For example, in a call center, there are task workers that are repetitively performing the same tasks over and over, and the job doesn’t really vary from person to person. Plus, they are generally working in a well-connected office. Therefore, terminal servers often works best in this situation. It provides the necessary connectivity; there is not much need for personalization of the desktops; and the shared CPU model is more than sufficient for the low performance demands. Clearly the cheapest computing model works well here.

In another example, a group of engineers may also work in a corporate office, so connectivity is not an issue, but they will not get the computing performance required to do their job in a shared terminal server environment. VDI (or even blades) might be a better option here because the desktop environment is dedicated to the individual (Note that blades provide a dedicated CPU and memory as well). There might also be some security reasons to use VDI or blades over distributed PCs, especially if outsourcing your engineering effort. We also have to acknowledge the road warriors and “work-from-home” groups that cannot count on a fat internet connection (or even a thin one in many cases), yet still must be productive. While they may still benefit from some forms of virtualization (application virtualization or streaming for example – see next blog), desktop virtualization is not going to be a good fit here. Physical systems may be more expensive, but are still a requirement for these groups to be productive. Remember that no organization will make money if their users cannot be productive – that’s what it’s all about.

So the quest for the perfect desktop environment is not really about TWO options – whether to virtualize or not, as much as it is to select the right balance of IT and end-user factors. To that end I will propose that there are really FOUR primary options to choose from for each of your end-user groups, each which deserve equal consideration:

• Terminal Servers
• Virtual Hosted Desktops (VDI)
• Blades
• Desktops and Laptops (yes, this could be 2 separate categories)

Can Desktop Virtualization save money? Yes, in some cases. Is money the only factor to consider? No, in all cases. It is a great thing to be able to standardize, and if you can get away with only 2 or 3 of these models in your environment, then great! But don’t compromise on productivity. Still, your total desktop support costs will be the sum of all models in place. And there are more costs to consider. Remember that all we have discussed here is where the working environment is, and how the user may access it. In the next installment, I’ll discuss Endpoint Virtualization, a collection of technologies that may help with the tools and assets within the desktop that the user needs to be productive, like applications, profiles, even the data itself.

Previous in the Series: I Know All About Virtualization, Don’t I?

Next in the Series: What is a Workspace?

PART 2:  Endpoint Virtualization Series

Many IT people today think they have a pretty good idea about what virtualization is, and even know some of the companies that sell it.  But perhaps the people that are confused have a better grasp on the situation, because they know that they don’t quite understand it all.  And that can potentially prevent leaping in a direction that does not provide the results you thought it would.  To make it more challenging, many vendors are bringing completely different architectural models to market, so it’s hard to stack rank “feeds and speeds” since in many instances you are evaluating fundamentally different approaches to solving problems.  And these approaches have enormous short and long term implications to all of your infrastructure, some that are yet to be fully vetted.  Let me see if I can help reduce the confusion by explaining a few things about the various virtualization technologies.

Let’s start with the one that I think people know the most about – Server Virtualization.  (I know this is pretty basic, but stick with me on this, and I think you’ll be better off in the end.)  The primary technology for this is called a hypervisor, and it is designed to abstract the hardware from the operating system.  By “virtualizing” the hardware, or creating a “virtual container” for the OS, one piece of hardware can be used for several operating systems, when only one would normally be allowed.  This has led to huge cost savings and greater flexibility in the datacenter.

Moving up the computing stack a bit, we come to Desktop Virtualization.  This is where a client operating system is abstracted from the local computing device, thin client or otherwise.  It just means that the OS is running somewhere other than where the user is sitting, usually in a datacenter.  Desktop Virtualization has actually been around in various forms much longer than Server Virtualization, going back to the mainframe days when people would access the system from remote terminals.  Now that desktop computing has been around for so long, many of us have forgotten and see this as revolutionary.  Although the concept is not new, there are some innovative new ways to accomplish it, with the main categories of solutions being various forms of “terminal servers”, “virtual hosted desktops” (also known as VDI), or blade PCs.  These are completely different technologies than Server Virtualization, and are adopted for very different reasons, like access security or hardware cost reduction.  It is also important to note that the Virtual Desktops (or remote OSes) may be running on Virtual Servers, or they may be running on physical servers – but the technologies are not dependent upon one another.

Now we get to Endpoint Virtualization, even higher in the computing stack, and perhaps the least understood.  Again, it is a collection of technologies that is sure to continue growing, but this time the purpose is abstracting the assets a user needs to be productive, like applications, profiles and data, from the operating system.  These technologies have more to do with end-user productivity by allowing applications and such to be independent of the operating system and device.  This allows a user to be more mobile and flexible, as his assets are not tied or dedicated to a specific environment or location.  Again, Endpoint Virtualization technologies, including application virtualization, application streaming, profile virtualization, and others, are completely independent of the previous two categories, and can exist in all environments, both physical and virtual.  For example, application streaming can help to consolidate application management, licenses and distribution, and deliver applications automatically to the right people, when they need them, regardless of where they are, or whether they are on a laptop in another country or a terminal server in the home office.

I believe that the greatest source of confusion today is a lack of understanding that Desktop and Endpoint Virtualization are separate solutions that solve different problems, and they should be evaluated on their own merits.  Often, companies that assume that virtualizing their desktops solves more problems than it does, wind up with a more complicated and expensive solution than is really necessary.  Likewise, Endpoint Virtualization does not address the location or security of the operating system itself, but its benefits extend to physical systems as well.

Tune in next time to look at how to go about selecting the right technologies to solve your biggest problems, without making them worse or more expensive.

Previous in the Series: Wow! I didn’t know Symantec did Virtualization

Next in the Series:   To Virtualize or Not to Virtualize: That is the Question

PART 1:  Endpoint Virtualization Series

If I had a nickel for every time I heard that phrase, I’d have a lot of nickels.  The point is that still an awful lot of people who are researching virtualization technology (because that’s what you do these days if you’re in IT) start with Citrix and VMware, perhaps go to Microsoft, then stop.  Certainly there are a ton of smaller companies being ignored that also have virtualization technologies.  But Symantec is a big company with virtualization technologies, and more people are learning about Symantec’s solutions every day.

Because Symantec holds such strong positions in security and storage, we are not normally the first company that comes to mind when thinking about virtualization.  However, there are some really good reasons that more and more companies, after looking at ALL of the options, are excited to select Symantec solutions.  In fact, the above phrase is usually followed by “This is great stuff!”

There are other reasons that people get confused about virtualization technologies, like not understanding the different types of virtualization, or which ones solve which problems, should they standardize on one technology, or is the best solution a mix of technologies.  Just understanding what technologies each company has to offer, and what they specialize in could help the investigative process a lot.  But there aren’t too many easy summaries or overviews to help with this.  I am going to try to do this for you.  You might think that this would be a biased account.  But interestingly, for the areas, where VMware and Citrix excel, Symantec doesn’t compete.  There is only a little bit of overlap, and we always encourage companies to do their own evaluations.  This allows us to work alongside these other companies quite happily, and I can honestly encourage a reasonable evaluation of the infrastructure alternatives.  And it will still benefit you to evaluate Symantec, regardless of your choices.

In case this sounds even more confusing, you are not alone.  Ask anyone and they will tell you the world of virtualization is a confusing place, and finding the right technologies to solve the right problems, without spending more money and adding more confusion to your environment, is an often unresolved challenge.  I’ll explain why that is and how to see the landscape more clearly, next time.

Next in the Series: I Know All About Virtualization, Don’t I?