Energy-aware Computing Continuum

Energy-aware service placement for Cloud, Fog, and Edge infrastructures. The repository contains a Python placement framework, GCP deployment helpers, and an Apache Storm integration layer.

Modules

• python_algo: placement algorithms, evaluators, properties files, and CSV exports.
• gcp_automations: GCP VM provisioning and Monitoring data collection.
• storm-scheduler: Apache Storm topologies and custom schedulers.
• scripts: helper scripts to build, start, stop, and submit Storm jobs.

Prerequisites

• Python 3.12+
• Java 17+
• Maven 3.9+
• Apache Storm 2.8.3 for cluster execution
• Google Cloud SDK for GCP deployment

Quick Start

Run a placement locally

cd python_algo
pip install -r requirements.txt
python main.py --strategy CSP --verbose

Run a placement with explicit properties

cd python_algo
python main.py --strategy GreedyFirstFit --infra properties/Infra_5nodes_GCP.properties --app properties/Appli_5comps_GCP.properties --placement-csv ..\results\placement.csv --metrics-csv ..\results\metrics_GreedyFirstFit.csv

Deploy a GCP infrastructure from an infra properties file

pip install -r gcp_automations/requirements.txt
python gcp_automations/deploy_gcp_from_properties.py python_algo/properties/Infra_5nodes_GCP.properties

Submit a Storm topology from an application properties file

./scripts/launch_topology_from_properties.sh ./python_algo/properties/Appli_4comps.properties DemoTopology

Run placement and topology submission together

./scripts/launch_placement_and_topology.sh ./python_algo/properties/Infra_5nodes_GCP.properties ./python_algo/properties/Appli_5comps_GCP.properties ./python_algo/properties/Infra_5nodes_GCP_mapping.csv CSP

Typical Workflow

1. Define infrastructure and application inputs in python_algo/properties.
2. Run a placement strategy with python_algo/main.py.
3. Export placement and metrics CSV files to results or results_infra10.
4. If needed, provision matching GCP VMs with gcp_automations/deploy_gcp_from_properties.py.
5. Build and submit the Storm topology with the scripts in scripts.

Outputs

• Placement mapping CSV: results/placement.csv unless overridden.
• Metrics CSV: appended to the path passed with --metrics-csv.
• Example benchmark outputs already exist in results and results_infra10.

Module Docs

• python_algo/ReadMe.md
• gcp_automations/README.md
• storm-scheduler/README.md
• scripts/README.md