Gartner and Microsoft make sure Linux numbers don’t add up. Counting, accounting, or discounting? Gartner and Microsoft make sure Linux numbers don’t add up.
I laughed out loud this past week when I read a new Gartner Group report that estimated Linux server “shipments” at just 8.6 percent of the total market. That number is a far cry from the numbers of other research firms, such as International Data Corporation, which estimates that Linux currently has nearly 30 percent server market share, with growth expected to rise to 38 percent by 2004.
Why the big difference in numbers? Well, first of all, the Gartner Group research received sponsorship from Microsoft, so that fact alone casts doubts in my mind about the accuracy of the numbers.
Next, interpreting Linux server market share accurately can’t be done by merely counting the number of servers shipped with Linux on them. Research firms traditionally analyze markets on this basis, but this type of analysis does not reflect actual activity when it comes to open-source software and its market share.
For example, one of my test networks has ten servers on it. When I purchased those ten servers, they came pre-loaded with Windows. They have since been wiped and totally reloaded with other operating systems, including Linux and FreeBSD.
If I used the metrics that research firms traditionally use to estimate market share, those ten servers would be counted as Windows market share, when in reality they are not running Windows at all.
The same is true on the higher end of the Gartner Group survey, wherein they estimated no Linux activity within the supercomputer sector. Here too, Linux is being loaded after product shipment. Moreover, companies also are loading Linux onto physical and logical partitions on midrange and mainframe systems. It’s not clear whether these post-shipment activities are reflected in Gartner Group’s numbers, either.
In short, you cannot use shipment as a means to accurately gauge usage of Linux or any other open-source software. It would be more accurate to sample the same number of firms in multiple markets (e.g., large enterprises, small businesses, ISVs, ISPs, ASPs, etc.), ask them for actual usage statistics, and then aggregate the overall numbers.
What should small-business leaders make of this report? It’s probably best viewed as an attempt by Microsoft to head off what is expected to be mushrooming growth in the use of Linux over the next few years. Presenting numbers that portray a “nobody is using it” result may keep many people from investigating Linux.
There is a larger point to take away from all of this as well. Research techniques that have been traditionally used to measure the computer industry in the past do not apply accurately to the computing marketplace emerging today, which includes open-source software that does not follow traditional shipping channel and revenue models.
Your best bet is to take any vendor-sponsored survey or research with a grain of salt. Look at all of the emerging as well as tried-and-true technologies, and then look at your business. Implement those solutions that solve your business problems in the most economical ways. After all, that is the point, isn’t it?
Contributing Editor Maggie Biggs has more than 15 years of business and IT experience in the financial sector.