The task-level and resource-level tools based on Regularity-based Resource Partitioning (RRP) model The current version is V0.9 (to be released). It includes the following tools.
PCG is the name used in the paper. It is actually the same thing as RSG. It was nameds so in the paper to better fit the context there. RSG takes in the information of a task set that will be deployed on a partition. It generates an availabilit factor (a resource specification) so that the task set given is schedulable on the partition with the availability factor assuming the task-level scheduling algorithm is Earliest Deadline First (EDF). The calculation is done based on theoretical analysis in [2].
CSG generates one cyclic schedule for each CPU with the configurations of the partitions (resource specifications) taken as inputs. It also takes the time-slice length as an input. The resource partitioning algorithm implemented here is MulZ-ILO, which is an approximation algorithm with proven theoretical bound. More details and analysis about MulZ-ILO is given in [1].
ST can judge whether the current configuration of a partition can offer real-time performance guarantees to the task set being deployed. The current version assumes the task-level scheduling algorithm used is Earliest Deadline First (EDF) and assume tasks are independent from each other inside the same partition. Calculation is implemented based on theoretical analysis in [2]. More complicated task schedulability tests introduced in [2] are yet to be implemented.
ReC-S is implemented based on the One-Hop Reconfiguration (OHR) model defined for embedded systems. It takes in a reconfiguration request, which includes the information of partitions to be scheduled and the scheduling deadlines of them. It also takes the time-slice length as an input. The algorithm implemented in OHR-OPT. The OHR model and algorithm OHR-OPT are introduced in [1].
Note that the current version only deals with Regular Partitions.
To run the application, you need to install Qt 5.12 or above. A packaged application without preliminary installs required will be added soon.
-
[1]: G. Dai, P. K. Paluri, A. M. K. Cheng and B. Liu, "Regularity-Based Virtualization under the ARINC 653 Standard for Embedded Systems," in IEEE Transactions on Computers, doi: 10.1109/TC.2021.3138019. Link: https://ieeexplore.ieee.org/abstract/document/9662228.
-
[2]: G. Dai, P. K. Paluri and A. M. K. Cheng, "Enhanced Schedulability Tests for Real-Time Regularity-Based Virtualized Systems with Dependent and Self-Suspension Tasks," 2021 IEEE 27th International Conference on Embedded and Real-Time Computing Systems and Applications (RTCSA), 2021, pp. 51-60, doi: 10.1109/RTCSA52859.2021.00014. Link: https://ieeexplore.ieee.org/abstract/document/9545293.