Internet has become an indispensable tool and a fundamental right, yet many regions, particularly in Africa, remain disconnected. Community networks low-cost wireless infrastructures managed by local populations have emerged as a viable solution to bridge this digital divide.
However, these networks, such as the one in Bibémi, face significant challenges. Located in remote rural areas, they are prone to hardware and software failures that can last for days due to limited accessibility. Anticipating these failures is crucial to ensuring network reliability and minimizing downtime.
Project Solution: CN Monitor
This work presents a specialized monitoring platform designed to proactively detect and alert administrators to anomalies within the Bibémi community network.
Technical Framework
The system leverages a modern data pipeline to transform raw network traffic into actionable intelligence:
Data Sources:
SNMP (Simple Network Management Protocol): Used to poll hardware devices (routers, switches) for operational parameters like CPU load, uptime, and bandwidth.
Log Files: Collected from network applications to identify software-level errors or security events.
Real-Time Processing:
Spark Structured Streaming: This Big Data framework processes incoming data streams in real-time, allowing for complex queries and immediate detection of threshold breaches.
Simulation & Testing:
The platform was validated using GNS3, where the specific architecture of the Bibémi community network was replicated to simulate realistic traffic and failure scenarios.
Operational Interface
A dedicated web application provides administrators with a global overview of the network health. Through a centralized dashboard, the system triggers immediate alerts when a failure is detected, allowing for rapid intervention even before the end-users notice a service interruption.
By shifting from reactive repairs to proactive supervision, this tool ensures that community networks in underserved areas become resilient and sustainable pillars of local development.