When developing a device or system, one of the earliest and most crucial decisions you need to make is deciding what type of operating system it’s going to run off. Linux is often the default OS of choice for many devices and projects, from Android smartphones and smart TVs to game consoles and automobiles.
However, while Linux and other traditional (or standard) operating systems run devices such as these that require intensive memory and processing power, RTOS (or Real Time Operating System) software is often chosen for devices where minimal latency is paramount, or where RAM is limited, to name a few scenarios.
Here at Nabto, as an IoT connectivity solution provider, we have plenty of experience dealing with both RTOS and traditional OS powered devices. In this article, we share our knowledge on RTOS v. OS, embedded Linux v. RTOS, and the advantages and disadvantages of RTOS.
Let’s firstly define exactly what an RTOS is and how it differs from a traditional OS. Put simply, it’s a software component that can execute only one program at a time, but which rapidly switches between programming tasks in order to execute multiple tasks “simultaneously”.
An RTOS delivers what’s known as deterministic, hard real-time responses to external events, which have to be completed within a specific time. Common examples of devices or systems that require this include anti-lock brakes on automobiles and driverless vehicle obstacle detection.
A traditional OS, in contrast, provides non-deterministic soft real time responses. It’s not hard to think of examples, given the multitude of applications you have running on your computer or smartphone at any one time.
Advantages of an RTOS
A common question we’re asked is “When should I use an RTOS over a traditional OS?”. While it’s difficult to advise on a device or system basis, as every project is different, it’s instead easier to outline the advantages of an RTOS.
- Priority Task Scheduling: Critical processing can be programmed to ensure repeated execution within the given timeframe/deadline. What’s more, critical processing can be seperated from non-critical processing tasks.
- Modularity/Defined Stack Usage: All RTOS platforms come with a mandatory core kernel, with customized middleware and protocol stacks. By building only what you need around the core kernel you’re able to control device memory requirements with predictable memory usage, which is much harder with a traditional OS.
- Reuse Code: Modularity allows for the buildup of a code base across devices, which makes development and deployment far more efficient and scalable.
- Efficient Memory Usage: While an RTOS requires both instruction memory (usually flash) and RAM for operation, the demands are far less than a conventional OS.
- Reduced Overhead: While an RTOS, of course, necessitates additional code beyond the application itself, it still often results in lower overhead and quicker responsiveness when compared to the alternative of using a polling loop.
- Cross-Platform Support: Many RTOS will be able to run on many different MCU platforms, making it possible to easily port an application from one to another.
- Abstraction of hardware: Many RTOSs will deliver a simple abstraction of some of the hardware, however, MCUs sometimes are so different that this isn’t always possible.
Disadvantages of RTOS
Of course, an RTOS also has specific drawbacks, and it’s important to consider these when assessing the choice between an RTOS or an OS for your project.
- Lack of Multi-Tasking: An RTOS is very efficient at managing a small number of scheduled tasks, but it’s much less efficient at managing system multi-tasking.
- Longer Wait for Low-priority Tasks: As an RTOS is programmed to execute priority tasks within specific deadlines, lower priority tasks may have to wait longer versus an OS.
- Minimal Task Capacity: As well as a lack of suitability with multi-tasking, an RTOS can only run minimal tasks simultaneously.
- Non-existent process and memory separation: An OS will be able to separate memory into different memory domains that can only be addressed by specific processes. This way, one program cannot “crash” another program. This is normally not a part of an RTOS.
Once you’ve assessed the advantages and disadvantages of RTOS v. OS for your project, you then decide on the best way forward. And if you’re deciding to use an RTOS, read our guide on how to choose the best RTOS for IoT.