Advice for venturing into virtualization

Server virtualization is all the rage, but to get the return on investment you're seeking, you've got to make a proper assessment. Gartner offers bet practices to avoid the problems that can plague server virtualization implementation.

“With the right virtualization approach and strategy, and with a long-term plan on the changes that virtualization will make to server management processes and tools (Compare Server Management products), enterprises will effectively leverage virtualization now and will re-architect their servers to become a more efficient, fluid pool of dynamic capacity,” says Thomas Bittman, vice president and distinguished analyst at Gartner.

Based on Gartner’s client advisory interactions, the research firm devised the following advice for embarking on a server virtualization project:

Start small, think big. Starting small is best from a cultural, cost and management perspective. Focus on server consolidation to reap cost savings and increased hardware use. The second phase shifts the focus to delivering new services or improving the quality and speed of service.

Require a rapid ROI. Show a full ROI within six months or less. Gartner recommends that firms deploying 50 or more virtual machines in a year should be able to make a good business case.

Virtualize the right apps. You’re best focusing on older, smaller packaged apps and not applications with high input output needs, which can be inefficient on virtual servers.

Set a storage strategy. Virtualization isn’t as useful if you store virtual images on direct-attached storage, because this limits your ability to replicate or recover those images. Plan to use a central storage system to give your organization flexibility.

Understand licensing. Vendors are still reacting to virtualization and pricing, licensing and support are in flux. See to understand these polices as thoroughly as possible.

Combine virtual machines effectively. The first step is consolidation, then you can worry about what goes where. Workloads change and the emphasis should be on dynamic relocation of server capacity.