In Anatomy of a Desktop Virtualization Project #1, we discussed getting to know your application estate and identifying the required delivery methods for those apps. In part #2, it’s time to engage with vendors and determine what form of solution you should deploy.

Why go through this stage? Surely you should just pick the best-of-breed solution and run with it?

Things are never that simple. First, the virtualization world is in a constant state of flux. Software vendors are continually expanding or condensing their solutions, adding and removing features almost incessantly. VMware and Citrix are two of the bigger players with ever-changing software portfolios.

Second, you can’t ignore the problem of cost. Unlimited budgets would be fantastic in all walks of life, but if that were the case, I’d be driving a Ferrari. The best solutions almost invariably have the biggest price tags attached to them.

It’s also true that at two different points in time, the best-of-breed solutions in a particular technological area may be completely dissimilar. I can recall a time when it seemed like App-V would become the de facto standard for application virtualization. However, in an evolving world, it now looks complicated and less agile compared to newer, simpler, leaner solutions, despite the fact that it still holds a healthy market share.

The skills required to design, build, and implement the solution are also a factor. If you do not have the skills in-house to do this, engaging with vendors directly, or with partners and consultancies who offer similar product portfolios, can help you past any lack of technical expertise by leveraging their professional services.

However, it goes without saying that vendors and associated partners/consultancies will always try to promise you the earth—particularly if they are marketing an end-to-end solution such as those provided by VMware and Citrix. End-to-end solutions can give you everything under one roof and even significantly reduce costs, but they also can introduce immaturity in certain areas of the solution, as well as vendor lock-in. In my experience, each use case is different. For most, a hybrid solution—involving two or more vendors—is usually the best fit. This is not to say that an end-to-end solution can’t be a compelling one, but a trade-off of sorts will occur somewhere within.

When engaging with vendors or partners in this way at the start of a virtualization project, it’s incredibly important to keep your eye on the ball and avoid being dazzled by new features. We touched on the importance of this in Setting Appropriate Desktop Virtualization Goals, but before you embark on the vendor engagement phase, you need to dramatically expand and define your requirements. Divide them up according to how vital they are to the project’s success. Engage with your users and support staff and find out what they don’t like with their current solution—what they’d like to see done differently. Once you’ve identified and decided on the list of requirements, don’t be tempted to shoehorn in other parts of the vendor’s proffered solutions just because they might be handy. If you’re already managing your user profiles through Group Policy/roaming profiles and the users and support staff are happy with it, why would you invest in a user profile management piece? If you’ve already decided that other requirements dictate your needs—for example, Citrix Platinum licenses—and then you get User Profile Manager thrown in for free, then that’s perfectly fine, but don’t get overwhelmed by a cool-looking demo and bring in software that increases your costs without ticking any of the requirements boxes.

Vendors and partners have a tendency to try to sell you a high-maintenance solution, for obvious reasons. The more moving parts it has, the more you will rely on them for help implementing and maintaining it. This is not to say that a complicated solution doesn’t have its place—if you have an extensive list of requirements and you need them all fulfilled, then you may have no choice—but overcomplexity can cause more problems than it solves. If you already had a solid OS deployment solution based around, for instance, SCCM, why would you purchase Citrix Provisioning Services? If there is no compelling reason to implement it that fulfills one of your requirements, it would be an inappropriate use of budget and resources.

On the flip side, it’s also important to take a longer-term view. Virtualization projects like this, particularly those aimed at “transforming the enterprise,” are often expensive and disruptive. It doesn’t make sense to perform a large programme of work in this fashion every few years, so it’s important to try your best to future-proof the solution. To this end, identify features and requirements that may crop up in the coming years, and build support for them into the solution, or at least leave the door open for it to be done without too much fuss. This can be difficult, particularly in industries that move into new areas, but staff tasked with taking a strategic view should be able to provide some input on this level.

Finally, if you don’t have any experience with the technologies the vendors or partners are pushing at you, it makes a lot of sense to bring in someone who does understand it well, even on a very short-term basis. Vendors and partners are not above a certain amount of downright subterfuge (or just incompetence). I can recall a consultant telling me that Unidesk would work on physical machines, which I had to very abruptly call him out on. Needless to say, the client I was working with was encouraged to avoid that partner at all costs.

You may come out of this phase happy to engage directly with a vendor or partner for an end-to-end solution, or you may decide that your needs are best served by a hybrid combination of software. There’s no right or wrong answer to the technology selection. The key is identifying the solution that ticks all the boxes you need it to and that can be implemented as smoothly as possible.

In summary, the vendor engagement phase can be broken down very simply:

  1. Know what you want.
  2. Know what they want.
  3. Know what they’re talking about, or bring along someone who does.

One reply on “Anatomy of a Desktop Virtualization Project #2: Vendor Engagement”

  1. I’d just like to add to this that it is vitally important to get the success criteria for each part of your project correctly defined at this stage. I’ve seen a lot of vendors come in, design a PoC, and then the PoC just seems to drag on forever, hamstrung by the lack of proper success criteria.

Comments are closed.