“Composability and Adaptability on a Time- and Space-Partitioned Architecture for Spacecraft Onboard Software”
Fast Abstract in the Doctoral Track of FACS 2010 - 7th Inter. Workshop on Formal Aspects of Component Software, Guimarães, Portugal, Oct. 2010.
Abstract: Space systems of the future demand for innovative computer architectures, enabling reduced size, weight and power consumption (SWaP) and component reuse among the different space missions. A solution being studied concerns the utilization of time- and space-partitioning (TSP) concepts, and fosters the interest of space agencies and industry partners. The AIR (ARINC 653 In Space RTOS) architecture was designed to fulﬁl the requirements for robust TSP and allow for mixed-criticality missions. Temporal partitioning is achieved through two-level hierarchical scheduling. In the ﬁrst level, partitions are scheduled cyclically. Inside each partition, processes compete according to the native process scheduler of each partition’s operating system. AIR implements the advanced notion of mode-based partition schedules, allowing temporal requirements (and, consequently, partition scheduling) to vary according to the mission’s phase or mode of operation. A formal system model of the AIR architecture was deﬁned in the course of the present work. We hereby address our current research work on TSP systems for aerospace applications, laid over the axes of published contributions on composability (and how it enabled component-based analysis of the system) and adaptability (self-adaptability, recon- ﬁgurability). We also introduce the forthcoming research direction of augmenting these properties with support for multicore platforms.
Research line(s): Timeliness and Adaptation in Dependable Systems (TADS)