Operating system/Kernel models

Introduction

edit

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

Design

edit
 
Graphical overview of a Monolithic kernel
  Search for Monolithic kernel on Wikipedia.

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.

Pros

edit
  • Speed
  • Simplicity of design

Cons

edit
  • 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

Examples

edit
  • Traditional Unix kernels (includes BSDs and Solaris)
  • Linux
  • MS-DOS, Windows 9x
  • Mac OS versions below 8.6

Microkernel

edit

Design

edit
 
Graphical overview of a microkernel
  Search for Microkernel on Wikipedia.

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.

Pros

edit
  • Stability
  • Security
  • Potentially more responsive (though often not in practice)
  • Benefits for SMP machines

Cons

edit
  • Additional context switches are usually required
  • Slow Inter Process Communication can result in poor performance
  • Not helpful for threads

Examples

edit
  • AmigaOS
  • Amoeba
  • Mach
  • Minix
  • MorphOS
  • L4
  • QNX

Hybrid kernel

edit

Design

edit
 
Graphical overview of a hybrid kernel
  Search for Hybrid kernel on Wikipedia.

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

Examples

edit
  • NT kernel (used in Windows NT, 2000, XP, Vista, 7, 8, and 10)
  • XNU (Mac OS X and ios' kernel)
  • DragonFly BSD
  • BeOS
  • Plan 9