redirect_from: /docs/concepts/

The use of Real-Time Operating Systems (RTOS) is a general practice in nowadays embedded systems. These embedded devices typically consist of a resource-constrained microcontroller that executes an application where the interaction with external components is performed. In many cases, this application contains a time-critical task where a time-deadline or deterministic response is required.

Bare-metal applications are also used nowadays, but it requires a very low-level programming skills and lacks of hardware abstraction layers that RTOSes offers. On the other hand, RTOSes typically uses hardware abstraction layers (HAL) that eases the use of hardware resources, such us timers and communication buses, making easier the development and allowing the reuse of code. In addition, they offer thread and tasks entities that, together with the use of schedulers, provides the necessary tools to implement determinism in the applications. The scheduling normally consists of different algorithms where the user can choose from. Another feature that RTOSes normally offers is the stack management, helping in the correct memory usage of the MCU, a valuable resource in embedded-systems.

Due to the benefits explained in the introduction, micro-ROS integrates RTOS in its software stack. The use of such a tool enhances the micro-ROS features and allows reusing all the tools and implementations they provide. As the micro-ROS software stack is modular, the exchange of software entities is expected and desired. Same happens with the RTOS. Even that NuttX is the *default* RTOS for the project, it is expected that several of them could replace it.