Skip to content
master
Go to file
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 
lib
 
 
 
 
 
 
 
 
 
 
 
 

README.md

Mixed-Criticality DAG Framework

License

This repository holds the Mixed-Criticality Directed Acyclic Graph (MC-DAG) framework. The tool is composed of three different components: a scheduler module for MC-DAGs, an availability estimator and a MC-DAG random generator.

General information

This framework uses a simple XML specification to handle Mixed-Criticality Systems. Some examples are available in btests and in the wiki.

The main propose of this work is to test scheduling methods that were developped during my thesis. By exporting modules into runnable JAR files, benchmarks can be performed. A python script to perform tests is provided in bench.

Scheduler

The scheduler uses a list-based heuristic to allocate tasks to cores. Static scheduling tables for all the criticality modes of a system are computed by this tool.

The scheduling tables for the high-criticality modes are computed first in order to determine safe activation instants for high-criticality tasks. The low-criticality scheduling table is then computed by allowing preemption and migration of tasks when the activation instants of high-criticality tasks occur.

Implementations of the global scheduling heuristic are based on G-EDF, G-LLF and G-EZL (EDF + Zero Laxity).

Random MC-DAG Generator

This framework includes a random generator of MC-DAGs. This generator aims to produce unbiased MC-DAGs to perform benchmarks on scheduling policies.

The unbiased generation relies on methods to create random topologies on DAGs, and uniform distributions of utilization to tasks.

Availability

This framework is also capable of performing model transformations to compute availability rates for low-critilicaty outputs of a MC-DAG. Models generated are probabilistic automata used by the PRISM Model Checker.

About

This implementation supports the following publications:

  • Directed Acyclic Graph Scheduling for Mixed-Criticality Systems - R. Medina, E. Borde and L. Pautet - presented at Ada Europe 2017.
  • Availability Enhancement and Analysis for Mixed-Criticality Systems on Multi-core - R. Medina, E. Borde and L. Pautet - presented at DATE 2018
  • Scheduling Multi-Periodic Mixed-Criticality DAGs on Multi-core Architectures - R. Medina, E. Borde and L. Pautet - presented at RTSS 2018.
  • Deployment of mixed-criticality and data driven systems on multi-cores architectures - R. Medina - Université Paris-Saclay (carried at Télécom ParisTech).

About

Mixed-Criticality DAG Framework

Resources

License

Packages

No packages published
You can’t perform that action at this time.