In the first episode of this step-by-step guide to virtualization adoption, I covered how to identify appropriate virtualization servers. Obviously, choosing and pricing servers must be done before ROI calculations can be made. Step two covered a delicate phase, capacity planning. In this phase, virtualization architects must carefully evaluate which kind of physical machines should host the desired virtual machines. This evaluation depends on specific arrangement strategies, and these strategies heavily depend on hosted applications workloads.
With these two steps completed, you have a good foundation of information to calculate ROI. It's time to detail these investments and compare them against the list of cost-saving benefits this technology provides.
To calculate ROI, you need to apply simple math to the costs your company could mitigate or eliminate by adopting virtualization. The operation is not trivial; some direct costs can be underestimated in the equation and indirect costs could be overlooked.
Virtualization can reduce some of the following direct costs:
- Cost of space (leased or owned) for physical servers
- Energy to power physical servers
- Air conditioning to cool the server room
- Hardware cost of physical servers
- Hardware cost of networking devices (including expensive gears like switches and fibre channel host bus adapters)
- Software cost for operating system licenses
- Annual support contracts costs for purchased hardware and software
- Hardware parts for expected failures
- Downtime cost for expected hardware failures
- Service hours of maintenance cost for every physical server and networking device
Some entries deserve a deep analysis, which should be reconsidered from project to project, given the incredible speed at which the virtualization market is changing. The cost of operating system licenses, for example, has not been a major factor until this year. Previously, customers deploying Windows had the same situation whether using a virtual machine or a physical server, but Microsoft slightly changed its licensing agreement this year in a way that facilitates virtualization adoption.
At the moment, Windows Server 2003 Enterprise Edition users can run up to four Windows instances, any edition, in virtual machines. The imminent Windows Vista will have a license that permits a second instance in a virtual machine. And the upcoming Windows Server codenamed Longhorn Data Center Edition will grant unlimited virtual instances of the operating system.
In the near future, it's highly probable that Microsoft will further revise this strategy allowing unlimited virtual instances of the host operating system as long as they remain on the same physical hardware.
In any case, all these direct costs strictly depend on the two factors we examined in the previous phase: VM/core ratio offered by the chosen virtualization platform, and most of all, the virtual machine arrangement we selected during capacity planning. But whatever arrangement or virtualization platform we use, some important indirect costs should be calculated as well:
- Time needed to deploy a new server and its applications
- Time to apply the required configuration
- Time to migrate to new physical hardware for severe unplanned failure
- Time to migrate to new physical hardware for equipment obsolescence
Although these factors cannot be easily quantified, they heavily influence the return on investment calculation for virtual infrastructures.
Speed and effectiveness in deploying new servers with pre-configured environments, something usually called capability to perform, is unbeatable in virtual data centers. In the same fashion, speed and effectiveness in moving the mission-critical applications from a failing hardware to a safe one, the so-called capability to recover, is a basilar feature of virtualization. No security solutions can compete with it.
ROI calculation is no easy task, and different interpretations could lead to very different results. To help potential customers, some vendors offer ROI calculation tools with partially precompiled fields, providing CPU/VM ratio values or product license prices. For example, SWsoft offers an online form for its OS partitioning product, Virtuozzo.
In other cases, skilled virtualization professionals have tapped their experience to create ROI calculation tools and then made them available to the whole community. Ron Oglesby, for example, has created an Excel spreadsheet called VMOglator to evaluate virtual machines' costs depending on physical hardware characteristics.
In most cases, these self-service calculators aren't 100% effective, because they can't track all factors in the cost-saving analysis, and because customers are often unable or don't have time to evaluate the real value of some expenditures. The best option would be a traditional ROI analysis, possibly through a neutral third-party firm, but that's a time-consuming and expensive service that few small and medium companies can afford.
At the end of this phase, we should create a list of costs for the current infrastructure (the so-called AS IS environment), a list of costs for the virtual infrastructure and a list of costs required for the adoption, including:
- The hours to recognize physical servers to virtualize
- The hours to complete the capacity planning
- The hours to calculate ROI
- The hours to migrate the actual infrastructure
- The hours and learning material cost to acquire required know-how for migrating the new infrastructure
- The software license cost for all tools needed in all adoption phases
These values are the final tools a company should use to understand whether or not virtualization is feasible or not and how soon it will pay off.
In the next part of this series, we'll enter the operational phase of the project, starting from the very first step: migrating existing physical machines into virtual machines through a process known as physical to virtual (P2V) migration.
This was first published in September 2006