Browsing by Author "Pinto, Joel Oliveira"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
- Mecanismos baseados em reservas para clusters de instrumentação de criticalidade múltiplaPublication . Pinto, Joel Oliveira; Maia, Cláudio Roberto RibeiroElectronics completely transformed the automotive industry as early vehicles were purely composed by mechanical components but the current reality is quite different. The growing acceptance for embedded electronics devices led to a significant increase in the number of microcontroller-based functions embedded in vehicles. With this increase, customer’s safety concerns raised. To ensure customers safety from the use of Electrical and Electronic (E/E) automotive equipment and systematic failures, Original Equipment Manufacturers (OEMs) and their suppliers must comply with standards such as ISO 26262, the road vehicles functional safety standard. ISO 26262 provides regulations and recommendations for the product development process. When the critical road functionalities are regarded as hard real-time, that shall complete within the defined time boundaries, coexist in an environment with soft and non real-time tasks (e.g., multimedia and connectivity activities) the system designer must use an approach to ensure that no critical activity is jeopardized in order to avoid hazardous events. To cope with the coexistence of activities with different time boundaries and criticality within the same system, this work proposes the implementation of uniprocessor reservation-based mechanisms, namely the Constant Bandwidth Server (CBS) and the Capacity Sharing and Stealing (CSS), in a real-time operating system for scheduling non-critical activities without jeopardizing the apriori guarantee of critical activities. Both schedulers use the concept of server, a task holder where a fraction of the processor bandwidth is reserved for tasks, thus relaxing the need for knowing certain properties of the tasks such as the WCET. Both implementations are detailed and compared through the implementation of task sets where both types of tasks coexist.