Strategic Desktop Virtualization
Desktop virtualization has become a very popular solution today and the virtualization ecosystem vendor community has responded with a wealth of products and solutions targeted at this market. Customers are generally overwhelmed and not well served by product marketing and vendors that are more interested in selling their products than they are in solving business challenges. The end result is often a product sale that may solve some of the customer challenges rather than focusing first on the customer’s needs and then providing a solution solely based on those needs.
A solid desktop design begins with an in-depth understanding and review of your business challenges, drivers, and the technical and non-technical issues that your teams are dealing with. The direction of the business, including planned growth and expansion, acquisition, or other major changes must be factored in when conceiving a desktop strategy.
Technical requirements also extend beyond normal design and integration efforts. Businesses need to evaluate their own internal technical capabilities and skill sets when embarking on a new desktop strategy. Often the skill sets required to design, implement, maintain, and support a new virtualized desktop environment extend past what a traditional desktop support and engineering team is trained and prepared for.
With this understanding of the existing environment, technology and required skill sets, a proper design can be built around the business’s specific requirements and not one focused on a specific vendor or technology. This forward thinking and the effort expended on the front side of a project can lead to great results on the implementation of the solution and potentially avoid unnecessary purchases of unneeded products.
Approaching the design of a desktop solution by focusing on the business requirements rather than on available technology can yield some interesting results. First, there have been and continue to be significant advances in virtualization technologies from many different vendors. By focusing on the requirements first, the value of these enhancements in relation to your specific requirements can be measured better. A clear understanding of the requirements will also help to focus your solution search on those that best address your needs.
In some cases, the challenge being addressed is more specific and less strategic in nature. Creating an alternative desktop design to address a directed or tactical problem is a great way to get started with desktop or application virtualization. This way a challenge can be addressed, ROI can be realized, and a solution with future implications for the rest of the environment can be put in place.
Keep in mind that even a tactical design requires a solid framework and plan to work from. Consider that decisions made on the tactical design and its implementation can be the foundation for future growth, and as such expansion should be considered a design element even where it cannot be accurately predicted at the time of the initial design.
The Dangers of Being Too Tactical…Stuck in the Weeds
One danger that is often faced by organizations delving into the desktop virtualization space is focusing too much on a specific use case. This could be a group of offshore developers who need better access into the network. The tactical solution is to put together technology from the vendors they are already working with to solve that specific challenge. Not much thought is put into the scalability of the solution, and even user-experience features are overlooked in an effort to quickly solve the specific challenge. In many cases, this is a short-term win for the group implementing the solution, however, if the tactical solution proves a success and initial confidence raised, more use cases for the technology emerge.
This is where the challenges begin to show themselves. The initial tactical solution is expanded and built on to the point where the management of it becomes a bigger burden than the problems it was initially intended to solve. Also, as the solution scales, some of the design elements and their costs to the organization begin to surface.
Another potential gotcha is when a solution meets a tactical need and is reviewed against a larger set of criteria and requirements. The tactical solution, which has the ability to meet or exceed the initial requirements, can be dismissed because it fails to meet the higher-level requirements. In this case, the business loses out on a potential win for their environment, and the project is either shut down or continues to revolve and evolve around requirements that may not be practically achievable.
Core to the challenge of scaling out the tactical design is the reality that desktop requirements and growth patterns are fundamentally different from the lessons learned in the server virtualization world. Virtualized servers are often deployed one at a time and for a specific purpose. This means that each server is unique and requires dedicated disk space and time to configure the application and services that will be hosted in a one-off manner.
Contrary to that design, virtualized desktops grow in packs. This means that the requests for desktops won’t normally come in one at a time, but can come in much quicker as use cases are defined and as new users are added to the solution. Even with the pain associated with virtual server sprawl, desktop sprawl has the potential to be many times greater and much more erratic.
The tactical solution design rarely takes this kind of growth or scalability into account. Available resources go on short rations as the desktop count increases, and the strain on server, storage, and network resources can begin to affect the user experience even further, reducing confidence in the solution. This approach also fails to take into account design elements that are not specifically related to solving the immediate challenge. For example, perhaps the immediate group of users has no requirements to view any graphically intensive content that may be a broader requirement for other use cases down the road. Other design issues that could be initially overlooked are those related to security or regulatory compliance. These have the potential to derail a solution pretty quickly.
Another example of a common tactical solution that has scalability ramifications is in the use of shared storage for the virtual desktops. Commonly, pilots and small production implementations begin with a simple approach to desktop virtualization by mimicking the “one virtual machine to one virtual disk file” design that is commonplace in server virtualization. This method is easily employed and demonstrates how effective a virtualized desktop can be. However, when the solution starts to grow and more desktops come online, the available storage capacity and performance quickly get overloaded and storage costs skyrocket while performance falls just as quickly.
In both of these examples, understanding and accepting that the solution may grow beyond the boundaries of the initial implementation will certainly lead to a more in-depth assessment and requirements-gathering exercise. By involving members and stakeholders from other business and technical units in your organization, the picture of what the solution could become will materialize and then the appropriate technology and design can be applied.