Operating system/Kernel models
The kernel is the name given to the program that does most of the housekeeping tasks for the operating system. The boot loader hands control over to the kernel, which initializes the various devices to a known state and makes the computer ready for general operation.
After preparing the hardware for use, the kernel provides an abstraction layer between the hardware and software running on the system. Through an application programming interface (API), software can request the kernel to perform hardware tasks, as well as access many common functions. Most current kernels also multi-task the hardware, sharing it between the various processes running on the system.
Monolithic Kernel Edit
Monolithic Kernels function with all of the kernel and device drivers inside one address space running in kernel mode. The principle advantage to this approach is efficiency (especially on x86 architectures or 32-bit computer systems where a task switch is a particularly expensive operation). Unfortunately, monolithic kernels suffer the risk of the entire system crashing due to a bug in a device driver, the reason being it is running in the same area with the kernel and other drivers, so, if something fails every other driver and program will be affected.
- Simplicity of design
- Potential stability issues
- Can become huge - Linux 4.15 has 20 million lines of code and Windows 10 contains over 40 million lines!
- Potentially difficult to maintain
- Traditional Unix kernels (includes BSDs and Solaris)
- MS-DOS, Windows 9x
- Mac OS versions below 8.6
A Microkernel tries to run most of its services and device drivers in userspace. This can result in an increase in stability and possibly security on machines with a Memory Management Unit. Unlike monolithic kernels, most of the components run in some kind of container, so if something fails, the module or driver will be restarted.
- Potentially more responsive (though often not in practice)
- Benefits for SMP machines
- Additional context switches are usually required
- Slow Inter Process Communication can result in poor performance
- Not helpful for threads
Hybrid kernel Edit
A hybrid kernel combines the concepts of both monolithic kernels and microkernels. It is generally implemented by having a monolithic kernel with a more microkernel like design. When properly implemented it is hoped that this will result in the performance benefits of a monolithic kernel, with the stability of a Microkernel
- NT kernel (used in Windows NT, 2000, XP, Vista, 7, 8, and 10)
- XNU (Mac OS X and ios' kernel)
- DragonFly BSD
- Plan 9