- of

10 Key Differences Between Desktop and Server Virtualization Deployments
by Chris Preimesberger
Predictable vs. Unpredictable Workloads
The load on a virtualized back end is quite predictable. While it's important to be able to move a server application from one hardware platform to another while runningthrough vMotion, for instance-desktops can be moved when the user is signed off and the desktop is idle.
Storage Is Handled Differently
Tier 1 storage for server applications often requires a high-speed SAN infrastructure to access data across disparate data stores. Tier 1 storage for virtual desktops can be simple DAS configured in server-based RAID, a less costly proposition.
GUIs Have Different Needs
Getting the GUI right is critical for a virtualized desktop user. Virtualized servers often have no direct interaction with the end-user community. The use of the SysAd command line is fine in the former, but an anathema in the latter. Ideally, end users would not need to know or care where the computing resource actually is.
Workload Differences
A virtualized server adds no appreciable networking load to the data center. A virtualized desktop operates entirely over the network. Getting the networking right is critical.
Integration Differences
A VDI must integrate seamlessly with the existing physical desktop infrastructure. Security, networking and authentication devices have all been optimized within budgetary, regulatory and compliance constraints. By contrast, server virtualization is a consolidation effort confined to the data center.
Organization Changes Differ
Desktop virtualization requires organizational change. The end-user support organization must function within the data center. Server virtualization requires an upgraded skill set, but the organizational structure is unchanged.
VDI Requires Changes to More IT Resources
Virtualized server performance varies depending on factors that are largely controlled within the data center. Virtual desktop performance can be affected by the data center, network and transmission protocol. In addition, diverse user requirements can also affect performance.
VDI Requires Changes to More IT Resources
Virtualized server performance varies depending on factors that are largely controlled within the data center. Virtual desktop performance can be affected by the data center, network and transmission protocol. In addition, diverse user requirements can also affect performance.
VDI Requires Changes to More IT Resources
Virtualized server performance varies depending on factors that are largely controlled within the data center. Virtual desktop performance can be affected by the data center, network and transmission protocol. In addition, diverse user requirements can also affect performance.
VDI Requires Changes to More IT Resources
Virtualized server performance varies depending on factors that are largely controlled within the data center. Virtual desktop performance can be affected by the data center, network and transmission protocol. In addition, diverse user requirements can also affect performance.
With IT managers keeping tight control over technology budgets as the economy struggles to recover, enterprises are taking a close look at virtualized desktop implementations, or VDIsknown a few years ago as thin clients. Server virtualization has been well-chronicled as a way to consolidate IT resources to create pools of storage and computing power and use it all to better advantage. Both types of virtualization have excellent efficiency attributes, are cost- and power-efficient when used correctly, and are centrally controlled. But that's where the similarities end.
Nonetheless, many people in the market still view server virtualization and desktop virtualization as having a similar architecture. In this slide show, we define the differences between the two types of virtualizations. Our information source is Leostream, a developer of virtual hosted desktop software, which provides the connection-broker apparatus required for enterprises to achieve useful large-scale desktop virtualization implementations.