Real-time systems for industrial use: requirements for the future
T. Gleixner · 2005
The current industrial use of real-time systems can be divided into two categories. High end systems, where commercial RTOS solutions have the main share, and low end systems, which are often driven by custom-built proprietary solutions based on a main program loop and interrupt handlers due to the traditionally restricted computing power of micro-controllers. The requirements for current industrial real-time systems are specific to the primary task of the device. The power of micro-controller platforms has increased significantly in recent years, making a more general software architecture for real-time controllers attractive. Increased power also increases the motivation to extend the duties of these controllers with a broad range of functionality, including non-real-time aspects not supported by traditional RTOS platforms. In addition, the range of applications with real-time components is expanding as well, with audio and video streaming being obvious but not unique examples. All of these trends are helping to erode the familiar distinction between real-time and non-real-time systems, increasing the need for a new approach combining the abilities of general purpose OS platforms with the low-level resource control capabilities of existing RTOS platforms. No existing platform can satisfy all of these requirements. Commercial providers are already extending the RTOS functionality by merging new functions including network connectivity, file systems, and graphic user interfaces into the real-time core. A different approach is the extension of a non-real-time OS with real-time functionality. WindowsCE is an example of an attempt to provide a limited set of real-time functionality on top of a multimedia centric general OS. Both of these approaches satisfy a limited audience because of their cost, and because they provide limited programming models. On the other hand we see a rapidly growing use of Open Source Software (OSS), especially Linux. The main benefit of OSS in these industrial applications is increased control over the system, along with scalability and flexibility; lower cost is a secondary attraction. the attraction of OSS for research is the ability to publish solutions. However, the available real-time extensions for Linux have some major disadvantages, which create a set of constraints that need to be satisfied by researchers addressing emerging industrial real-time requirements within an OSS platform: