Skip to content

Threat Detection System

Welcome to the central documentation hub for the Threat Detection System (TDS) β€” a high-availability, edge-computing threat detection sensor network deployed across a multi-node Raspberry Pi cluster.


πŸ—ΊοΈ Documentation Map

Navigate through the guides and configurations structured by component and layer:

βš™οΈ Cluster Infrastructure & Setup

Guide Description
❗ (MUST FOLLOW!) Cluster Shutdown and Cold Start Guide Step-by-step shutdown and boot recovery operations.
Cluster Setup V2 (Setup & Verification) Verification checklist and system checks for network boot.
Ansible Orchestration Guide Setting up ssh keys, hosts inventory, and batch task automation.
Basic User Guide Day-to-day admin commands and Ansible cheat sheet.

πŸš€ Distributed Web Services

Guide Description
Local Dev & Workstation Setup Run the entire FastAPI + Databases stack locally using Docker Compose.
Distributed k3s Kubernetes Deployment Scheduling API replicas, node affinity rules, and Ingress routing.
MinIO S3 Object Storage Guide Configuration for distributed S3 buckets and erasure coding over NFS.
Edge Camera Test Stream Guide Ingestion scripts, base64 image streaming, and browser test client.
Camera Stream β†’ Database Guide YOLO detection pipeline, MQTT topics, and how events reach PostgreSQL.
πŸ“‹ Changelog v1 β†’ v3 All backend, frontend, edge node, and K8s changes made in the v3 update.

πŸ“Š HPC Benchmarking & Parallel Computing

Guide Description
MPI Cluster Setup Guide Cross-compiling OpenMPI via chroot and launching cluster-wide jobs.
HPL Synthetic Benchmarks Tuning process grids (P x Q) and running LINPACK benchmarks.
Task Distributor Guide Validating Amdahl's and Gustafson's Laws using parallel POV-Ray renders.

πŸ“‹ Project Task Matrix

Detailed completion progress for each task requirement:

Task Title Description / Status
Task 1 Edge Computing Infrastructure Network boot setup, dual-homed DNS config, and SSD migration.
Task 2 HPL Performance High Performance LINPACK benchmarking and process tuning.
Task 3 MPI Cluster OpenMPI deployment and distributed workload validation.
Task 4 Non-MPI Scaling Laws Scaling analysis (Amdahl vs Gustafson) with POV-Ray & Task Distributor.
Task 5 Monitoring Prometheus, Grafana, and system health status.
Task 6 Model Training Custom YOLO model training and threat detection weights.
Task 7 Backend API Load-balanced FastAPI replicas with MinIO S3 and Postgres integration.
Task 8 Frontend Dashboard React web console displaying node stats and live threat streams.
Task 9 Telegram Notifications Telemetry bot alert integration.

Cloud Computing Course SS2026 β€” Frankfurt University of Applied Sciences
Project Repository: Threat-Detection-System