This repo contains the partitioning algorithm, the comparison simulator and the simulated results of the scientific paper titled 'Real-time task scheduling in a FaaS cloud', presented in IEEE CLOUD 2021
First, you need to fine tune the simulation parameters:
- TASK_COUNT (number of tasks 'n')
- CPU_COUNT (number of processors 'm')
- SIMULATION_COUNT (simulation repeat count with input 'n' and 'm')
To configure the parameters, just open the file simulation_for_paper.py in directory cloud2021_paper_materials and modify the variables at the begining of the source code.
The simulation results will be placed into csv files in the same directory from where the simulator was executed.
Run simulations:
cd cloud2021_paper_materials
./simulation_for_paper.py
We have created jupyter notebooks to process the simulator output csv files and generate easily readable plots to conclude the results. The notebook files are located in the directory cloud2021_paper_materials:
- Evaluation_heterogeneous_multicore_multinode_cluster.ipynb
- Evaluation_homogeneous_multicore_multinode_cluster.ipynb
The simulated data from the paper are located in directory cloud2021_paper_materials/dataset_paper.