With server virtualization being all the rage, it can be very tempting to jump into it with a "build it and they will come" mentality. This could be risky, as recent surveys have indicated that a sizeable number of adopters aren't able determine if their projects were successful. We shouldn't forget that a virtualization project is no different than any other large scale IT undertaking: it takes careful planning, clearly defined objectives, and reliable execution in order to realize the benefits. Here are a few items to help avoid some common pitfalls:
1. Quantify projected cost savings with straightforward, easy-to-quantify metrics. You can't determine project success without a yardstick, but stick to hard dollar savings and avoid soft, fluffy or complex TCO/ROI calculations.
2. Prepare to virtualize a substantial number of servers. Realizing the value depends, to some degree, on scale. Develop a formalized target server list and ensure that resources and commitments are in place to virtualize them.
3. Ensure application owner buy-in. Application owners don't like to feel like guinea pigs and may raise concerns over potential service impact. Have solid resource utilization metrics in place and ready answers for common concerns (e.g., performance, availability). Strong executive sponsorship will also help. And, of course, always have a back-out plan.
4. Create a plan for ongoing operational process integration. This is where much of the "heavy lifting" comes. Processes for monitoring and managing virtual machines must be integrated into existing workflows, and it will be necessary to enhance standard procedures and possibly acquire additional tools to better support virtualization.
5. Re-evaluate capacity planning and resource requirements. Virtualization makes provisioning easy -- sometimes too easy -- and it's possible to overrun server and storage capacity sooner than expected.
6. Enlist the right cross-functional resources in the project process. This is not just a server consolidation, it's a major infrastructure change and the other segments of the infrastructure supply chain, including storage and networking, need to fully participate.
Jim Damoulakis is chief technology officer at GlassHouse Technologies. He can be reached at firstname.lastname@example.org