infoTECH Spotlight Magazine

infoTECH Magazine

January 01, 2011

The Ghost in the Machine: Calculating Virtualization's Hidden Costs

By TMCnet Special Guest
Jesse Lee, Co-Founder, Senior Director of Technical Marketing at Apptio.

This article originally appeared in the Jan. 2011 issue of InfoTECH SPOTLIGHT

Virtualization is the tide that lifts all boats. In the past five years, enterprises have embraced virtualization with gusto to reduce their data center footprint and optimize their IT infrastructure. Most of these new virtualization initiatives began innocently enough, starting within the closed sandboxes of testing and development. Now, as virtualization has proven its value, it has begun to creep into enterprise production environments with the pace and momentum of Japan’s bullet trains. But there’s still a long way to go with Gartner (News - Alert) estimating that fewer than 15% of production servers are virtualized. The reasons for this slow adoption cycle are manifold and well known: security concerns, the complexity of managing virtual environments, and policy compliance, to name just a few.

However, one of the biggest obstacles hindering the adoption and acceleration of virtualization projects is often one of the most misunderstood: the lack of visibility into the underlying costs coupled with an inability to accurately calculate ROI from these initiatives. When a physical server is dedicated to a single IT service, allocating the cost of a server is a relatively simply process. However, when that same physical server is virtualized and provisioned out to many services, it becomes significantly more complex to track infrastructure costs and manage cost allocation policies. And when those virtual machines are dynamically moved between server pools, costs become even more difficult to track.

So essentially what you have is a perfect storm of abstraction whereby the underlying costs of all these IT resources is constantly in flux. For IT leaders, this makes it all the more difficult to build solid, data-driven business cases to justify new virtualization initiatives. Nor can they rely on assumptive data to make their case that virtualizing a pool of servers will pay off in the long run. Rather, IT needs to speak the language of the business to justify these initiatives. The building blocks of this language are financial metrics that can be clearly articulated to business stakeholders. Only then can IT declare with confidence that a given virtualization project also makes good fiscal sense.

Allocating Costs in the Virtual Realm


Building a cost allocation model is the first step in understanding the fully burdened cost of both your physical and virtual infrastructure. This model represents in aggregate all of the underlying costs that comprise an operating environment – everything from the physical servers, networking, storage, software licenses, and support staff to the power and facilities themselves. Until all of these costs are modeled, it’s simply not possible to understand your true costs. There are several accepted allocation strategies that can be used to attribute costs. Some of the more popular ones are as follows:

·         Even Spread - Dividing IT costs evenly among virtual machines, applications or services is the easiest way to perform cost allocation. With this approach, IT cost data is simply split into equal parts. This allocation is easy, but not fair as it has no representation of how much resource a service consumes.

·         Manually Assigned Percentage – This method provides more accurate cost assignment than the Even Spread methodology. With this allocation strategy, someone in the company who can provide an educated guess of how costs should flow will assign appropriate allocation percentages.

·         Indirect Weighting - This allocation method leverages a third data source to allocate the weighted percentage of a shared resource. For instance, virtual machines might be allocated based on memory assigned. You could create a percent allocation based on the percentage of memory allocated to each virtual machine. Similarly, storage can be assigned to applications based on allocated space.   

·         Activity Based Costing (ABC) – This method is more sophisticated and the most accurate. Activity Based Costing tracks IT activity that actually happened, as captured in a system of record, then uses those numbers to distribute shared costs. For example, help desk costs are allocated according to per-ticket costs driven by help desk users or infrastructure maintenance. 

These are just a few examples of the types of cost allocation strategies that can be implemented to get a grip on your true infrastructure costs. As any managerial accountant will attest, you can’t hope to understand your ROI until you know what it all costs.

Jesse Lee is the co-founder and senior director of technical marketing of Apptio.

TMCnet publishes expert commentary on various telecommunications, IT, call center, CRM and other technology-related topics. Are you an expert in one of these fields, and interested in having your perspective published on a site that gets several million unique visitors each month? Get in touch.

Edited by Stefania Viscusi