Step 1 — Assess your server environment.
A number of valuable tools on the market today enable the collection of performance metrics from your server environment. These tools can help you identify which servers would make good virtualization candidates. A partial list of available tools includes:
o Capacity Planner
o vCenter Server Guided Consolidation
o Microsoft Assessment and Planning Toolkit (MAP)
o System Center Virtual Machine Manager
o System Center Operations Manager
Step 2 — Finalize candidates for virtualization.
Before making any final decisions about which servers to virtualize, it’s important to understand each software vendor’s support policies. Is the software supported to run in a virtual machine? Is it supported to run on specific platforms (e.g., Oracle, Microsoft, etc.)? It’s also critical to determine each software vendor’s licensing policies.
• Does the policy allow for virtualization of the software?
• Does it permit dynamic movement of the virtual machine via vMotion,
XenMotion or other means?
• If the host is licensed, does the virtualization policy allow for additional
software instances to run at no additional cost?
Once you’ve answered these questions, you will be able to build a final list of virtualization candidates.
Step 3 — Determine the virtualization platform.
Citrix, Microsoft and VMware produce some of the well-known virtualization platforms, or hypervisors. Each platform is different. Once you’ve identified the features, functionality and caveats of each hypervisor, you can make an informed decision about which one will best meet your specific business needs.
Step 4 — Determine the server hardware platform.
To realize the greatest cost savings, determine if server reuse is an option for your business. Note, however, that any system more than three to four years old will probably be removed from newer virtualization platform hardware compatibility lists (HCLs). Warranty expirations and renewal costs also represent a key consideration in deciding on the merit and cost-effectiveness of server reuse. Finally, some hypervisors require specific processors with specific features (AMD-V or Intel VT) to operate properly. The hypervisor you select may, therefore, determine the feasibility of server reuse.
If new server hardware is the right option, consider rack servers, blade servers, AMD and Intel options and embedded hypervisors. HP, IBM and Sun all manufacture server hardware that leverages virtualization technology.
Step 5 — Determine storage hardware platform.
EMC, HP, IBM, NetApp and Sun Microsystems all produce leading storage systems. Again, it is important to identify the features, functionality and caveats of each platform in order to determine which one will best match your particular business needs.
Step 6 — Revisit backup/restore architecture.
Many vendors offer enhanced backup products for virtual infrastructure. Some of these vendors price by socket or host, thereby reducing costs considerably. If your business is tied to a long-term maintenance contract, consider finishing it out before changing your backup architecture. For those ready to investigate backup/restore architecture, the following companies offer some of the best options for virtualized infrastructure:
• CA • Commvault • EMC • IBM • Microsoft • Symantec • Vizioncore
Step 7 — Understand server operating system licensing.
If you plan on running many Windows Server instances, consider licensing each host processor with Windows Server 2008 Data Center Edition. Doing so enables unlimited Windows Server instances with no additional charge as well as dynamic movement of instances from host to host. These benefits prove extremely useful if you plan to clone and test Windows Server instances. Linux distributions have similar offerings but will vary by vendor.
Step 8 — Plan the project.
Assigning a project manager remains key to ensuring a successful virtualization implementation. Working with the project manager, you should build a project plan.
Determining the vendors through whom to acquire the necessary hardware, software and services represents another critical aspect of planning. Procuring everything from a single vendor will often yield the deepest discounts. But be sure to research all possibilities.
Finally, make sure to build a conversion time line — and stick to it. This step will prove one of the most effective means of controlling cost.
Step 9 — Educate and Implement.
For the virtualization process to remain cost effective, you’ll need to educate your implementation team. This education could take place in various ways:
• Online, classroom or onsite training
• Onsite jumpstart engagements
• Planning and design workshops
The virtualization platform should be rolled out simultaneously with the server and storage hardware. Once everything is in place, you’re ready to optimize the environment, which should include documentation and testing of backup and, if applicable, disaster recovery procedures.
Step 10 — Leverage physical-to-virtual (P2V) conversion to the greatest degree possible.
To realize the greatest cost savings and ROI, you’ll want to eliminate as many physical servers as possible through virtualization. You’ll also want to achieve this conversion quickly and efficiently. After accomplishing these conversions, don’t forget to turn off old systems. These systems can be either recycled, repurposed or sold.
With fewer physical servers, don’t forget that you’ll have less heat dissipation. Maximizing your savings will thus also involve a reassessment of your data center cooling distribution.
As a final consideration, it may be most cost-effective to outsource P2V projects. Utilizing a team of experienced P2V engineers can save a substantial amount of time and money.