Data center space, power, and cooling Network bandwidth Security Manageability Disaster recovery Image and license management
See “Step 3: Analyze Total Cost of Ownership” for more information on calculating infrastructure requirements.
Best Suited For Environments where application and license deployment and management are critical. With application streaming and virtualization, IT can use policies to control when licenses expire, whether the PC is connected to the network or not. Example: setting the license to expire for temporary or contract employees when their contracts have completed. IT can then repurpose the license. This method of delivering applications is additive to any of the other image delivery models. Applications can be streamed to any operating system, whether it is traditionally installed, virtualized in the data center or on the PC, or streamed. Endpoint Considerations While this model can be used with either thin clients or intelligent clients, using an intelligent client has some advantages in mobile usage scenarios. With an intelligent mobile client, users have “offline capability”— they can still access applications resident in the cache even when they can’t connect to the network. Benefits Security: Applications are stored and protected in the data center; local corruption is removed and patches are updated at each initiation from the streaming server; isolating applications limits data exposure. Manageability: Application licensing and provisioning can be streamlined through centralized management; application virtualization may enable legacy applications to run on a newer operating system; reduced conflicts, corruption, and randomness in the operating system registry.
Performance: Virtually identical to locally installed applications; good computing/ graphics-intensive performance; after launch, network demand drops to very low levels; reduced network demand versus streaming the operating system; user experience is the same as a standard PC boot. Infrastructure cost: Fewer/less costly servers needed; streaming technology has the lowest initial deployment costs of centralized computing models. Disaster recovery and business continuity: Users can shift to a different site or client; users can continue to work offline with cached content. Mobility: Streamed applications can be cached for off-network use on mobile clients. Limitations Security: At runtime, data and applications are vulnerable to client-side attacks or theft. Performance: Streaming download speeds can be affected by distance from server, network load, and number of users; interactions between virtualized/nonvirtualized applications can be affected. Software sequencing process: Not all software is suitable; initial sequencing setup/debugging can be time and labor intensive; streamed and virtualized application interactions can be problematic.
Application Streaming and Virtualization
How It Works Client-based model. The client operating system is locally installed, but applications are streamed on demand from the data center to the client, where they are executed locally. Streamed apps frequently do not install on the client operating system, but instead interface with an abstraction layer and are never listed in the operating system registry or system files (hence the term “application virtualization”). Interactions between streamed applications, other locally installed software, and the operating system are minimized, eliminating software conflicts. It also can improve security by “sandboxing” applications in isolated containers. Streamed applications can be cached on a laptop and taken off the network. When reconnected, the application can resynchronize with the server to check licensing, version, and usage information, and download new application data to the data center.
11 Intel IT Center Planning Guide | Desktop Virtualization