Real World Open Source: The TCO Question

Wednesday Aug 24th 2005 by Aaron Weiss

TCO is often a key decision factor when deciding between an open source or commercial solution. How the amount is measured, however, is more important than the amount itself, and should reflect the organization's priorities.

How much is this going to cost me? The question may be the lifeblood of business, but the answer is all too often an alchemy of knowns and unknowns, hard figures, and soft projections. Occupying the gap is a cottage industry of studies, surveys, analysts, and pundits weighing in on the big, bad, broad comparison of total cost of ownership (TCO) between commercial and open source software.

Some suggest that less than 30 percent of an organization's software TCO is in upfront costs (e.g., sale price or licenses), while the rest is in support and know-how. One study has argued that the typical Linux TCO may be half that of Microsoft Windows, while one Microsoft-sponsored study suggested Windows offered an 11 percent to 22 percent savings compared to a similar Linux-based deployment. These and the myriad studies like them, nearly all rooted in one agenda or another, reveal virtually nothing about your organization's needs.

In the real world, these seemingly comprehensive studies wind up either so broad or so scenario-specific that they are more useless than the analogous dietary allowances printed on food labels. A hypothetical average American might require 300 grams of carbs per day, but how much do you need? Like making dietary decisions, the TCO question must be evaluated in the context of an organization's own specific metabolism.

Upfront Costs and Internal Synergy

Commercial software is ostensibly designed to meet the needs of customers, which may include administrative facilities. In contrast, open source software is typically designed to meet the needs of a task.
Perhaps the most widely known fact about open source software is that it is usually available at low to no cost. You can go online and download Apache, the world's most popular Web server, freely and instantly. But is Apache the most popular Web server because it's free, or because it's the best, or both? Is soccer the world's most popular sport because it requires the least investment (one ball) to play, or because it's the best, or both?

Apache is a great example of how internal synergy can benefit open source software. The initial product met a need. A core group of people adopted it because, whatever its limitations, it was free. Its easy availability led to increasing popularity which, because of its open source collaborative development model, led to rapid product evolution and enhancement. The more widely Apache was adopted, in part because it was free, the better the product became, which further promoted wider adoption. This synergy has led Apache to power more than 70 percent of Web servers worldwide.

But not all open source projects enjoy this level of success. Defenders of commercial software argue that the upfront cost savings attributed to open source are illusory because of the support and labor costs that follow. That argument is strongest when applied to open source projects that haven't yet attained internal synergy. It is weaker when applied to those that have.

Labor vs. Licenses

Most critics of open source TCO point to "hidden" labor and support costs. Described broadly, commercial software is ostensibly designed to meet the needs of customers, which may include administrative facilities. In contrast, open source software is typically designed to meet the needs of a task. Commercial software makers must ultimately generate a profit to justify the venture. Open source developers ultimately aim to please themselves and satisfy their own needs, without regard to market share.

On the face of it, the open source methodology seems like bad news for business customers. But the story is more complex than that. When initially implementing open source software, it is reasonable to expect to invest in labor and knowledge — human capital — rather than fees or licenses. It may also be true that an initial software investment may favor commercial software, when licenses are less expensive than people.

The interesting trend, however, is a potential shift in costs over time. Commercial software vendors must generate ongoing revenues, which often leads to license renewal fees and enforced upgrade programs. Commercial vendors may also drop support and compatibility for applications as they age. Often these costs are controlled by the commercial vendor and its needs ׫ which means an organizations will likely be forced to upgrade applications because of others' requirements.

Open source software demands only an investment in human capital, a cost that can be controlled by the organization itself rather than a third party.

Open source software demands only an investment in human capital, a cost that can be controlled by the organization itself rather than a third party. While myriad studies debate which approach costs more or less in absolute terms, one fact remains clear: Vendors pull the strings when it comes to commercial software costs, both in terms of dollars and cents and timing. Open source software, in contrast, allows organizations to be in charge of their own expenditures.

Critics of open source costs emphasize that the model lacks "someone to lean on," asking "Who do you go to if you have a problem with the software?" But really the question is, "Where you want your money going?"  — to outsourced experts working on behalf of a commercial vendor, or in-house experts, who are employed to solve your organization's individual needs.


Commercial vendors such as Microsoft have been lowering some of their per-seat and per-CPU costs to better compete with the open source zero-fee temptation. For a fixed-size organization, per-seat pricing alone may not tip the scale toward open source. But what about an organization with significant growth plans? Whether it's desktop employees or back-end servers, those per-seat and per-CPU costs can add up.

It doesn't cost any more for an employee to set up a four-CPU server than a single CPU server. But most commercial software agreements would require four licenses to run on that machine; an open source license retains a cost of zero.

Thus, for organizations that want to better control the costs of their own growth, the TCO of open source software can scale well.

Crunch the Numbers

There is no one-size fits all answer to the TCO question. Forget about ideology and whether open source developers share the same profit motive you do. The crux of the comparison between commercial and open source software is where the better human capital investment lies — in a third-party vendor, or in your own experts.

For short-term or task-specific or turnkey solutions, TCO may lean in favor of commercial vendors. For long-term customizable and adaptable solutions, TCO may lean in favor of open source products, and keeping long-term cost control in-house.

Mobile Site | Full Site