An elastic mixed-criticality scheduling framework for cyber-physical systems




Su, Hang

Journal Title

Journal ISSN

Volume Title



In recent years, the rapid growth of cyber-physical systems (CPS) attracts the research interests from both industrial and academic communities. With increasing needs of executing tasks with multiple-critical levels on a shared computing system, scheduling mixed-criticality tasks while satisfying their specific requirements has been identified as one of the most fundamental issues in CPS. Most existing mixed-criticality scheduling algorithms guarantee higher level worst-case execution times (WCETs) of high-critical tasks at the expense of discarding low-critical tasks, which can cause control system to suffer from significant performance loss. To provide minimal service guarantee for low-critical tasks and stabilize real-time control system, an Elastic Mixed-Criticality (E-MC) task model and the associated early-release EDF (ER-EDF) scheduling algorithm are proposed. In ER-EDF, low-critical tasks are allowed to be released at least once during their maximum periods (i.e., minimal service level) while ensuring the worst-case timing constraints of high-critical tasks are always met. During run-time, slack time generated from high-critical tasks can allow LC tasks to release more frequently, which improves their control performance. The ER-EDF are studied on single and multi-core system. Observing the overheads associated with run-time slack management for early-releases, the E-MC model is extended to allow each low-critical task have a pair of small and large periods, which represent its service guarantees in low and high running modes, respectively. The dynamic-priority (DP) and fixed-priority (FP) scheduling algorithms are proposed for the extended E-MC model (E-MC 2), where their schedulabilities are analyzed with demand-bound function (DBF) and WCET response time analysis (RTA) techniques, respectively. In addition, period selection and priority assignment are also investigated for optimizing control performance of real-time CPS.


This item is available only to currently enrolled UTSA students, faculty or staff. To download, navigate to Log In in the top right-hand corner of this screen, then select Log in with my UTSA ID.


control performance, mixed-criticality tasks, multi-core system, real-time scheduling



Computer Science