But Linux also presents a unique set of management challenges that differ markedly from operating systems such as Windows or Unix. Massive heterogeneity of distributions, lack of nimble management tools, and the sheer volume of equipment and data all mean headaches for administrators trying to run an enterprise data center on Linux. Gartner research estimates that 74% of all server failures are caused by human errors including incorrect software configuration, and that over 50% of the time application upgrades and rollouts fail due in part to a lack of sophisticated management tools for Linux.

Take customization as an example. For most operating systems, there usually isn’t much variation in the OS. With Linux, there are scores of different distributions that can each be customized at installation to perform a different task. This high customizability of Linux is an upside, of course, because it allows for greater flexibility in the data center. The downside, though, is that a data center with many different OS configurations on different server boxes requires a huge amount of expensive administrative effort, labor, and time.

Due to this and other Linux-specific issues, the expense of server management in a Linux-based data center can actually offset most of the cost savings associated with commodity hardware. In fact, some statistics show that the cost of managing servers running Linux can be four times the cost of the hardware itself – and can account for 60% of the total cost of running a data center. Add in the opportunity costs associated with fixing mistakes caused by human errors, and that number can grow even higher. The prohibitive cost of managing large and complicated installations over time, making sure every single server (up to thousands of servers) is running correctly, has been an obstacle to Linux achieving its full potential in the data center.