Computing systems are becoming more and more complex and powerful every year. It is nowadays not uncommon to run several server applications on the same physical platform. This gives rise to a need for resource reservation techniques, so that administrators may prioritize some applications, or customers, over others. This article gives a brief introduction to the Linux kernel 2.6 task scheduler. The article also presents an implementation of a scheduling mechanism, that in a non-intrusive way introduces CPU bandwidth reservations for a task, or a group of tasks, in the GNU/Linux operating system. The scheduling mechanism is first tested using ded icated load tasks, and then on a setup consisting of two Apache servers.