microkernel

A microkernel is a minimal operating system kernel that provides the most essential functions and runs in privileged mode. In a microkernel design, core services such as interprocess communication (IPC), basic scheduling, interrupt handling, and address space management are implemented in the kernel, while higher-level services like device drivers, file systems, and network protocols run as user-space processes called servers. This separation aims to improve modularity, fault isolation, and portability.

The architecture centers on a small, portable kernel that relies on IPC to communicate with user-space servers. By keeping most services outside the kernel, a microkernel can be simpler to audit and to formally verify, and failures in servers do not crash the entire system. However, this design commonly requires efficient IPC and context switching to reduce overhead.

Microkernel concepts arose in the 1980s with work by Jochen Liedtke and were implemented in systems such as Mach (CMU), QNX, and later MINIX 3 and L4 family variants. SeL4 is a formally verified microkernel. Real-time and embedded systems have adopted microkernels for safety and reliability.

Advantages include improved reliability and security through isolation, easier updating and customization, and potential portability across platforms. Disadvantages include higher performance overhead due to more user-space communication and more complex design to achieve parity with monolithic kernels in terms of device support and driver integration.

Modern microkernel implementations focus on IPC efficiency, capability-based security, and formal verification. They are used in embedded, real-time, and secure systems, and have influenced hybrid designs that blend microkernel features with monolithic drivers.