Debunking the “No ROI in Virtual Desktop Computing” Myth

By Julian Weinstock
published: Tuesday, August 26 2008

Virtualization has been the hottest topic dominating the technology headlines in the last year, whether it is about server virtualization initiatives or benefits, or about new developments around virtual desktop computing. In fact, Gartner expects virtualization to be the highest-impact trend in IT infrastructure through 2012.

The leading edge of that trend is server virtualization, but virtual desktop infrastructure (VDI) is rapidly catching up based on its promise to reduce the complexity and cost of enterprise desktops, and change the way we work,  However, as is the case with many emerging trends, VDI economics is still subject to different assumptions and interpretations.

The early returns hold that there is no ROI in desktop virtualization. It's a wrong, but understandable, assumption.

Server virtualization, which has already gained significant market traction, offers compelling ROI based on server consolidation.  The higher degree of asset (server) utilization translates directly into considerable reductions in server capital investment (CAPEX), but has only marginal impact, if any, on the operational management costs (OPEX).  The success of server virtualization has created the expectation that VDI would provide similar ROI, based on CAPEX savings.  However, as some VDI skeptics have noted, this is not the case.

The skeptics' argument that there is no ROI in VDI is based on a back-of-the-envelope cost calculation along the following lines:  $400-$500 per desktop investment is required for server, storage, network and other data center costs to set up the VDI infrastructure, and $100-$200 is needed for a thin client or other access device.  The total cost of $500-$700 per client is too close to the cost of a rich client to provide compelling ROI for VDI, even if the useful lifetime of VDI is longer than the typical three years of distributed, rich desktops.

The fallacy of this argument is simple: the challenge with enterprise desktops today is not the acquisition costs but the total cost of ownership (TCO), according to Gartner, IDC and other leading industry analysts.  Acquisition costs, which have been tracking a declining trend for a long time, represent less than 20% of enterprise desktop TCO.  However, desktop management costs are rising faster than acquisition costs are declining, and represent more than 80% of the TCO.  So, even if VDI did reduce desktop acquisition costs, it would be meaningless when the TCO is defined predominately by the management costs.

As the skeptics correctly point out, VDI does not reduce the acquisition costs.  However, it does slash the management costs quite significantly-and this considerably reduces desktop TCO.

Management matters

VDI promises a number of benefits to the enterprise, including enhanced (physical) security and better regulatory compliance; higher availability and built-in business continuity -- with the desktop available anytime, anywhere, on any device; increased business agility  -- with the ability to provision or move a desktop with a click of a button; reduced power consumption; and remarkably better manageability.  Of these benefits, better manageability has the greatest impact on reducing the cost and complexity of the desktop and, therefore, the highest impact on TCO.

Enterprise desktop TCO numbers vary but, in general, industry analysts peg the cost at $3,400 - $5,900/desktop/year, depending on the extent of management, and the type of the client device.  About half of the TCO is attributed to direct IT cost of managing the lifecycle of the desktop, and the other half is attributed to indirect costs associated with lost user productivity.  The high management cost is driven primarily by the distributed and heterogeneous nature of enterprise desktops and the inseparable coupling of the desktop image to both a user and a physical device (and its device drivers).  The majority of direct costs stem from two labor-intensive management activities: helpdesk services and IMAC (installation, moves, adds and changes). The rest come from more automated activities such as security and compliance monitoring, asset inventory, and software distribution, patches and updates.

VDI Saves 15-25% Over Traditional Desktop Environment

VDI is a centralized, homogeneous and device-independent model, with the desktop image completely abstracted from and independent of the underlying device or infrastructure.  This architecture is inherently more manageable and, in fact, it can effectively reduce the top two cost line items (helpdesk and IMAC) by more than 50% each.  Software distribution, patches and update components of the TCO can also be reduced by employing the pooled desktop option of VDI. Even though, to capture these gains, upfront capital investment and on-going operational expenses for the VDI infrastructure is required, in total, VDI can reduce desktop TCO by about 15%-25%.

This is quite a compelling proposition, especially when extended to thousands or tens of thousands of desktops, but it is not without challenges of its own. Acquiring and operating VDI infrastructure (server farm, storage array, network gear, virtualization platform, etc.) of any meaningful size introduces a new dimension of complexity and cost, placing VDI economics out of reach of many organizations, both large and small.  

Extending the ROI Benefits of VDI

Imagine another computing model that could capture the desktop management savings and reduce OPEX without any upfront investment in infrastructure, and even without the burden and expense of operating the infrastructure.   Similar to the SaaS model, whereby the infrastructure is owned and operated by a third-party and provided as a subscription service, the VDI infrastructure could be owned and operated by a third-party and provided as a service on a subscription basis.

Such a desktop as a service (DaaS) model already exists and is radically changing the economics of VDI.  It enables enterprises to capture the benefits of VDI without having to own and operate the VDI infrastructure.  This transforms CAPEX to OPEX and can save enterprises another 10% over internally deployed VDI, effectively saving 25%-35% over traditional desktop practices.

As you can see, the skeptics are partially right:  VDI does not result in clear CAPEX savings like server virtualization does.  But, let's not ignore the savings it does deliver in desktop manageability and operational expenses.  That ROI is considerable, and it can be made even more dramatic by consuming VDI as a service.

Desktone , The Evolution of Corporate Computing: From SaaS to DaaS