How your commands get to the tape drive.
On one end is an application, like tar. This is the command that a user will normally invoke to access their tape drive. How does this finally result in an action at the tape drive itself?
Any application that is designed to make a peripheral do something has to operate something like what is shown below.
App ---> Kernel ---> driver #1 ---> ... ---> driver #n ---> peripheral.
The application programmer includes code that indicates they want a particular peripheral to do something. When the program is run this request is passed to the kernel, as it has final authority over who gets to use the peripheral and under what circumstances.
The kernel determines exactly what peripheral is under discussion. Based on that it passes the request to another piece of software called a driver. This software is designed to make that particular hardware operate. It may be that there are several drivers who will be invoked, one by the other, as each decides that another driver must be invoked to handle the request appropriately.