How pfandzelter/optimal-leo-placement works
Project Type and Purpose
• This is an open-source research tool (Python 3.9+) for QoS-aware resource placement on Low Earth Orbit (LEO) satellite edge computing networks.
• It provides:
• A configurable simulation environment (simulation of satellite constellations and inter‐satellite link distances).
• A placement algorithm for deploying edge services to meet Service Level Objectives (SLOs).
• Post-processing modules for analysis, summarization, and visualization of results.
File-Structure Analysis
Top-level scripts & modules:
• config.py – central configuration of simulation parameters, shells, SLOs.
• distances.py – entry point to run constellation simulations and produce ISL distance data.
• placement.py – core placement algorithm (QoS-aware resource placement).
• shells.py – wrapper to execute placement.py over configured shells/SLOs.
• placement-distances.py – analysis script combining placement output with distances to evaluate SLO compliance.
• summarize.py – aggregates experimental results into a CSV.
• graphs.py – generates visualizations from summary CSV.
Subpackage:
• simulation/ – code based on SILLEO-SCNS, including animation.py, constellation.py, simulation.py; used by distances.py.
Other:
• requirements.txt, Dockerfile – environment definitions.
• LICENSE files – MIT for main code, GPL-v3 for simulation subpackage.
Key Insights from README
• Heavy use of standalone CLI scripts; no GUI or external service calls beyond Python libraries and optional Docker.
1. Configuration Manager (config.py)
2. Constellation Simulator (simulation/ + distances.py)
3. Placement Engine (placement.py + shells.py)
4. SLO Analysis Module (placement-distances.py)
5. Summarizer (summarize.py)
6. Visualizer (graphs.py)
7. Data Storage (CSV/JSON output files: distances.csv, placement results, results.csv, graph images)
8. Environment Layer (Python runtime + Docker container)
• Configuration → Simulator → distances.csv
• Configuration → Placement Engine + distances.csv → placement_output.json
• Placement Output + distances.csv → SLO Analysis → analysis_output.csv
• analysis_output.csv → Summarizer → results.csv
• results.csv → Visualizer → graph.png / HTML
c. Architectural Patterns:
• Pipeline pattern (discrete stages connected by data artifacts)
• Modular package structure (simulation submodule separated from core logic)
• CLI‐driven orchestration (each script is an entry point)
d. Technologies & Frameworks:
• Python 3.9+, standard libraries, NumPy/Matplotlib (in requirements.txt)
• SILLEO-SCNS codebase integrated under simulation/
• Docker (Dockerfile) for reproducible environment