Energy-Efficient Fault Tolerance for Real-Time Tasks with Precedence Constraints on Heterogeneous Multicore Systems

Overview

This repository provides a simplified Python implementation based on the paper:

Energy-Efficient Fault Tolerance for Real-Time Tasks with Precedence Constraints on Heterogeneous Multicore Systems
Abhishek Roy, Hakan Aydin (George Mason University), and Dakai Zhu (University of Texas at San Antonio)

It implements DAG-based task modeling, basic partitioning (LTF, TBLS), speed assignment (Uniform Scaling, CPSS), and fault-tolerance via contingency tasks. While not a full production-grade system, it demonstrates core ideas in a concise form.

Features

1. DAG Task Model
   • Each task has a base execution time, references to successor tasks, and a frequency parameter.
2. Heterogeneous Dual-Core Architecture
   • Simulates a high-performance (HP) core and a low-power (LP) core.
3. Partitioning & Ordering Heuristics
   • Largest Task First (LTF): Sort tasks by descending execution time; assign to the core with the lowest utilization.
   • Threshold-based List Scheduling (TBLS): Assign tasks to the LP core until the utilization exceeds a threshold, then assign to the HP core.
4. Speed Assignment
   • Uniform Scaling (US): Same frequency for all tasks.
   • Critical Path-based Static Speed (CPSS): Detects a simplified “critical path” (in this demo, tasks with the largest execution times) and boosts their frequency.
5. Fault Tolerance
   • Each task is duplicated on the opposite core, executed only if a primary fault is detected.
6. Energy Model
   • Simplistic formula ( P(f) = a \cdot f^3 + \alpha ), integrated over execution time.

File Structure

├── README.md 
└── energy_fault_tolerance.py 

• energy_fault_tolerance.py: Main code containing classes (Task, Core), DAG generation, partitioning heuristics, speed assignment, duplication for fault tolerance, and a simple main function that simulates example configurations.

Installation and Usage

1. Clone or Download this repository.
2. Python Environment: Any Python 3.7+ environment should work. It is recommended to use a virtual environment.
3. Install Dependencies: If a requirements.txt is provided, install via:

   pip install -r requirements.txt

python energy_fault_tolerance.py