JOIN OUR
COMMUNITY

TRY OUR SOFTWARE

Virtual Desktops

The expression desktop virtualization is related to multiple technologies (Terminal Services, Remote Desktop Services, VDI, Application Virtualization, DaaS…)

One of the main reasons why the number of VDI projects isn’t growing at the speed predicted by experts is that some IT departments aren’t capable of identifying a business problem for which the already said technology is a valid solution

Most IT departments have only undertaken server virtualization projects that have been a complete success for the organization itself in relatively recent time period, which would make it logical to think that the next natural step would be to undertake a desktop virtualization project. Unfortunately, in most cases they don’t take into account that these are completely different projects

Virtual desktops, different technologies

Facing the challenge of outlining any desktop virtualization process, it helps to know the different desktop virtualization technologies and platform management methods that are currently available

Server-Hosted Virtual Desktops. These are known as SHVD (Server-Hosted Virtual Desktops)

It consists basically of a desktop operating system hosted on a virtual machine found in the Data Center, with each user having his or her own desktop. These desktops can be one of two different types:

  • Persistent. In this type of desktop, all changes made in the O.S., applications and user profiles are saved each time the user closes his session. These desktops allow users to personalize their desktops in great detail, including personalization for applications installed by the users themselves. This is the closest thing to traditional physical desktops. When you use persistent virtual desktops, you lose a large part of the advantages and usage philosophy of virtual desktops, since the problems with traditional platforms are transferred onto the new platforms deployed
  • Non-persistent. The virtual desktops deployed return to a state defined by the template from which they are derived once the user closes his session or the administrator makes changes on the template.These types of desktops require a system in which the user profiles are not associated with the virtual desktop, as all user configurations would be lost. In the same way, data is stored on external devices

In both cases, these virtual desktops can be accessed via LAN or WAN by using different remote protocols such as RDP, Remote FX, RGS, HDX or PCoIP

Client-Hosted Virtual Desktops. These are known as CHVD (Client-Hosted Virtual Desktops). Two types of hypervisors are involved in these platforms:

  • Type 1 (native, or bare metal). These types of hypervisors are installed on the hardware, incorporating their own drivers to make resource requests to the available hardware directly
  • Type 2 (Hosted). In this case, the hypervisors make resource requests to the O.S. on which they are installed 

In a CHVD environment a template is created in the Data Center and is then deployed in the different client devices. Subsequently, this image is synchronized in the Data Center once the device is connected to it

These types of environments are aimed at client devices that move through areas in which connectivity with their Data Centers is practically non-existent

Presently, this technology brings about serious problems with applications connected to database servers and with the synchronization of user profiles 

Terminal Services.  Multiple users open different sessions on a single operating system and can show the entire desktop or just the window of the application selected by the user 

Desktop as a Service (DaaS). It can’t be considered a technology in and of itself. Rather, it’s more of a pay model based on the consumption of resources of the desktops used. The service provider becomes the one responsible for maintaining the user stations (O.S., applications, user profiles, etc.), thus reducing the platform operating costs (Opex). The technology used will be whatever the service provider has decided to implement in their rented infrastructure

Application virtualization and user profiles

In both cases, we’re dealing with two terms that can be used in both traditional user station platforms and in virtual desktop environments. Occasionally, this is misinterpreted as desktop virtualization

Application virtualization.This was initially used to isolate applications that were incompatible in a single O.S. (for example, applications that use different versions of Java)

One advantage of application virtualization is the possibility of streaming the application, or using the application according to demand, without having to perform any kind of installation on the O.S. underneath it

In most cases, this technology is used thanks to the ease of deploying on-demand applications in accordance with departmental criteria

This technology allows IT departments to deploy applications based on layers: 

  • Corporate applications. They are installed together on the virtual machine’s O.S., which is then used as an image to deploy the virtual desktops
  • Other applications. They are deployed dynamically using application virtualization according to departmental or temporary criteria (developmental groups, training environments, etc.)  

In different environments, application virtualization can be very difficult to implement, particularly in proprietary applications and interrelated applications that share data with each other

User profiles. This methodology is based on isolating user profiles so that they can move between different desktops. This type of implementation is fundamental in non-persistent virtual desktop environments due to the aforementioned functioning thereof 

Future of desktop virtualization

Desktop virtualization is a mature enough technology for its adoption to result in multiple advantages, both for the user and for the administrator of the platform

It allows us to optimize exploitation costs and take full advantage of our resources, obtaining higher levels of security, flexibility and speed when carrying out deployments, which are attractive enough terms for information departments and CIOs, not to mention CEOs and financial departments

SSD disks. Using these types of disks, even in limited way, in shared storage reduces the probability of performance issues in virtual desktop platforms. The cost of this type of storage technology is getting lower and lower and produces extremely visible results regarding performance
 
GPU virtualization.  This technology is used for high-performance virtual desktop environments (graphic work stations, engineering stations, etc.). It can be implemented in two ways:

  • By providing the possibility of directly assigning physical GPUs to virtual machines
  • By using specific graphic cards which, together with a specific hypervisor, generate a pool of virtual GPUs ready to be assigned to the virtual desktops 

Collaborative virtual desktop platforms. VDI desktop platforms tend to be systems that support and manage practically any service that a user can access or use (remote storage, multimedia storage, remote sessions on physical devices, etc.). These services can be interrelated among each other, forming “deployed services” sets available to the end user

Additionally, independent work groups or freelance developers will be able to incorporate new “access” or “service” models to these types of platforms that will automatically become available to the users

Open Source desktop. The choice of the O.S. will be much more open than it is nowadays. The O.S. of connected devices or of the virtual desktops themselves could be Linux, MacOS, Windows or any other system, with users being able to access the corporate applications from any of those. The criteria for using the O.S. will be determined by the adaptation of each user and work station, performance, features and costs

In organizations without special requirements, or in those that have followed a strategy of “freedom” with their infrastructures, systems with Linux and open code applications will become more and more common, as they will be able to obtain all of the features they need at zero acquisition cost. New collaborative applications based on the Internet will also incentivize the use of this architecture

Read the complete white paper here