AnTi-MiCS: Analytical Framework for Bounding Time in Embedded Mixed-Criticality Systems
New framework balances worst-case execution time to improve utilization and Quality-of-Service
In mixed-criticality (MC) systems, tasks have different criticality levels, and the Worst-Case Execution Time (WCET) serves as a conservative upper bound. However, using a single low WCET for lower-criticality modes is challenging: too low risks frequent mode switches hurting QoS, too high wastes utilization. Behnaz Ranjbar and Akash Kumar present AnTi-MiCS, a design-time analytical framework that selects the optimal low WCET by analyzing task execution patterns. This approach achieves a 30.27% improvement in QoS and a 35.89% reduction in utilization waste compared to existing methods, as demonstrated on real embedded benchmarks.
For cases where a single low WCET fails (e.g., bimodal execution distributions), the authors extend their work with MulTi-MiCS. This scalable method computes multiple low WCETs by leveraging temporal correlation between successive inputs. MulTi-MiCS further enhances QoS by 6.41% and reduces utilization waste by an additional 8.23% over AnTi-MiCS. The paper, published on arXiv (2604.27862), targets embedded real-time systems and offers a practical trade-off between resource efficiency and system reliability.
- AnTi-MiCS analytically determines optimal low WCET for mixed-criticality embedded systems, improving QoS by 30.27% on average.
- The framework reduces utilization waste by 35.89% compared to existing approaches, balancing efficiency and reliability.
- MulTi-MiCS extends the method to handle multimodal execution distributions, boosting QoS by 6.41% and cutting waste by 8.23% further.
Why It Matters
Better time bounding enables safer, more efficient embedded systems in automotive, aerospace, and IoT applications.