Endpoint Considerations This delivery model can be used with both thin clients and intelligent clients. Thin clients are most appropriate when users don’t need to be mobile; when no user personalization is needed or it can easily be removed by reimaging; and when the applications being accessed are not computing- or graphics-intensive. Intelligent clients are most appropriate when rich content and computing- or graphics-intensive applications will be pushed to the client. Before making the thin or intelligent client decision, take into account what type of applications and content users will require not just today, but also two to three years into the future. Benefits Security: The operating system, applications, and data never leave the data center. Manageability: Desktop image and data management are centralized for simpler administration, application management, validation and support, and more reliable backup; there are reduced image management and storage challenges if using public images instead of private images. User customization: Allows a more PClike experience; personalization is possible if private images are used. Access: Users can access applications from any network-connected client.
Hardware/software image validation: A single “platform” target for all operating system and application images reduces validation efforts. Disaster recovery and business continuity: Users can shift to another client or site. Limitations Performance: Perfomance degrades as the number of users per server increases; there can be graphics bottlenecks; the model requires a continuous stream of low-latency bandwidth to maintain display, keyboard, and mouse responsiveness. Manageability: IT must manage, update, and patch virtual desktop images stored in the data center. Software and peripheral compatibility: Not all software or specialized peripherals are compatible or suited to this approach. Mobility: VHD requires a persistent network connection with adequate bandwidth. Costs: New deployments are expensive, including data center space, servers, software, networking, and storage; this is the most server-intensive delivery method. Disaster recovery and business continuity: If server function is lost, processing must shift to a redundant server or data center; loss of network function renders clients inoperable. User satisfaction: VHD is unlike the PC experience in performance, customization, flexibility, and mobility.
ISV Products Citrix XenDesktop Microsoft VDI Mware View
Application Level or Image Level?
A helpful way to think of the models and how they fit with your requirements is whether the problem needs to be solved at the application level or image level. In this case, an “image” is the complete package of the operating system, applications, and user data and settings. Some compute models solve application problems, while others solve image problems. Application-level models are: Terminal services Application streaming and virtualization Image-level models are: Virtual hosted desktops Operating system image streaming Client-side virtual container For example, an application-level model is likely best suited for a “bring your own computer” implementation, in which end users use their own personal computers at the workplace. In this case, delivering an entire operating system image is likely unnecessary, and streaming only the applications the user needs is sufficient. If an intelligent desktop is used, this approach has the added benefit of enabling users to access their applications even if they can’t connect to the network, because the applications remain in cache until users can log in again.
Intel IT Center Planning Guide | Desktop Virtualization