Cognit Optimizer

A decoupled daemon that optimizes device-to-cluster assignments using Mixed-Integer Linear Programming (MILP) to minimize energy consumption and carbon footprint.

Overview

The optimizer periodically:

1. Reads device assignments from the local SQLite database (device_id, cluster_id, app_req_id, estimated_load)
2. Fetches each device's application requirements from OpenNebula using the app_req_id
3. Filters feasible clusters based on device constraints (flavour, confidentiality, providers, geolocation)
4. Runs MILP optimization to find optimal cluster assignments that minimize energy/carbon
5. Scales clusters in parallel based on optimizer output via cluster scaling endpoints
6. Updates the cluster_id in the database for devices assigned to successfully scaled clusters

Database Integration

Local Database (device_cluster_assignment.db):

• Stores device-to-cluster assignments with application requirements
• Shared between cognit-frontend and cognit-optimizer
• Schema: device_id, cluster_id, flavour, app_req_id, app_req_json, estimated_load, last_seen

cognit-frontend interaction:

• When a device requests a cluster, cognit-frontend queries OpenNebula for feasible clusters
• Selects the closest cluster and stores the assignment in the database
• The optimizer later re-evaluates these assignments globally to optimize across all devices

cognit-optimizer interaction:

• Reads all device assignments from the database
• Uses app_req_id to fetch fresh requirements from OpenNebula
• Optimizes assignments considering all devices simultaneously
• Scales clusters in parallel via cluster scaling endpoints
• Updates cluster_id in the database incrementally as clusters finish scaling

Dependencies

Required:

• cognit-frontend repository at /home/ubuntu/cognit-frontend/ (imports db_manager.py and opennebula.py)
• Database file at /home/ubuntu/cognit-frontend/database/device_cluster_assignment.db
• OpenNebula API accessible at http://localhost:2633/RPC2
• OpenNebula credentials in ~/.one/one_auth or /var/lib/one/.one/one_auth

Setup

1. Ensure cognit-frontend repository exists:

   ls /home/ubuntu/cognit-frontend/src/db_manager.py
   ls /home/ubuntu/cognit-frontend/src/opennebula.py

2. Create virtual environment:

   cd cognit-optimizer
   python3 -m venv venv
   source venv/bin/activate

3. Install dependencies:

   pip install -r requirements.txt

Usage

Run single optimization cycle:

python main.py

Run as daemon (continuous mode):

# Use default interval (300 seconds)
python main.py --daemon

# Use custom interval
python main.py --daemon --interval 60