Subscribe vs. build

Two weeks ago I noted that, although the adoption of open source solutions with commercial support is growing (think Red Hat and SpikeSource), the price/performance ratio of those solutions remains excellent, even if you’re spending a little on support. I also pointed out that open source has practical limits and mentioned that InfoWorld is planning to retire its open source implementation of SpamAssassin in favor of an outsourced solution.

A number of readers wrote in to say their experience with open source has been similar to mine -- not perfect, but mostly rock-solid with a few operational issues. But one reader posted a deconstruction of my column on a Linux message board, concluding that my decision to outsource spam filtering simply meant that I did not understand how e-mail worked and therefore I had no right to offer advice on IT, period. “You cannot write coherently about it if you don’t use it,” the post read. When you really boil it down, though, successful IT is all about good judgment. My specific knowledge of the problem at hand and the available solutions made it abundantly clear that it would be an incredibly foolish business decision to do anything but outsource the problem.

Anyone who reads this column or has worked with me knows I’m an enthusiastic implementer of open source solutions. But as I mentioned in my earlier column, open source for me has never been as much about the politics of open source as it has been about meeting an IT need in the most effective way possible, in terms of both cost and functionality. The recent emergence of a class of highly functional, highly scaled, and reasonably low-cost subscription-based services means that perhaps for the first time, certain types of outsourced services can handily beat open source implementations in overall cost.

In the case of the outsourced spam solution, I did the math and realized that the monthly fee for the outsourced service was roughly equivalent to one hour per month of our very capable system administrator’s time, not to mention that we wouldn’t require servers, backup tapes, power, and cooling. Functionally, the outsourced service also offers a better end-user UI to manage personalized spam rules. For me, the choice was a no-brainer.

The operations expertise and hardware required to run an open source solution aren’t free and never will be -- and even the most brilliant systems administrators and developers have limited amounts of time. Fortunately, seasoned IT managers know that when you carefully pick the right outsourcer, you can relieve pressure on staff time and sometimes solve the problem at hand for a lower actual cost than if you took the DIY approach. The spam example aside, consider CRM: An open source shop might be tempted to try duplicating the functionality of, say, Salesforce.com or Salesnet using the ever-capable LAMP (Linux Apache MySQL PHP) architecture. But do the math, and in most cases, it would cost substantially more to try to build and maintain a less capable system than to just subscribe to the service.

The old adage that “time is money” has never been more true, and as outsourced services grow less expensive and more capable, the business justification diminishes for keeping IT functions such as spam filtering and CRM in-house, even when running on “free” open source platforms. IT shops that resist even considering such services are committing the cardinal sin in business: wasting both time and money.