Server rental store

AI in the Isle of Man Rainforest

AI in the Isle of Man Rainforest: Server Configuration

This document details the server configuration supporting the "AI in the Isle of Man Rainforest" project. This project utilizes artificial intelligence to monitor and analyze data collected from our unique rainforest environment. This guide is intended for new system administrators and developers joining the team. It assumes a basic understanding of Linux server administration and networking concepts. See Help:Contents for basic MediaWiki editing guidance.

Project Overview

The "AI in the Isle of Man Rainforest" project aims to leverage sensor data – temperature, humidity, sound, and visual feeds – to understand the rainforest's ecosystem. We employ machine learning models to identify species, detect anomalies (such as unusual sounds indicative of distress), and predict potential environmental changes. The entire system relies on a robust and scalable server infrastructure. More information on the project goals can be found on the Project Homepage.

Server Architecture

The infrastructure is composed of three primary server roles: Data Acquisition, Processing, and Presentation. These roles are physically separated for redundancy and security. Each role is detailed below. For details on our Security Protocols, please see the dedicated security documentation. Server access is managed via SSH Key Authentication.

Data Acquisition Servers

These servers are responsible for collecting data from the deployed sensors. They are located close to the sensor network to minimize latency. They perform initial data validation and buffering before transmitting the data to the processing servers.

Server Name Role Operating System CPU RAM Storage
rainforest-daq-01 Data Acquisition Ubuntu Server 22.04 LTS Intel Xeon E3-1220 v6 16 GB DDR4 1 TB SSD
rainforest-daq-02 Data Acquisition (Backup) Ubuntu Server 22.04 LTS Intel Xeon E3-1220 v6 16 GB DDR4 1 TB SSD

These servers use a custom Python script, detailed in the Data Acquisition Script Documentation, to interface with the sensors via Modbus TCP. Data is transmitted over a dedicated VLAN to the processing servers.

Processing Servers

These servers are the heart of the AI system. They receive data from the acquisition servers, run the machine learning models, and store the analyzed results. They require significant computational power and memory.

Server Name Role Operating System CPU RAM GPU Storage
rainforest-proc-01 AI Processing CentOS Stream 9 2 x Intel Xeon Gold 6248R 128 GB DDR4 NVIDIA Tesla V100 4 TB NVMe SSD
rainforest-proc-02 AI Processing (Backup) CentOS Stream 9 2 x Intel Xeon Gold 6248R 128 GB DDR4 NVIDIA Tesla V100 4 TB NVMe SSD

The machine learning models are built using TensorFlow and PyTorch, as described in the Machine Learning Model Repository. The processing servers utilize a distributed computing framework (Kubernetes) for scalability and fault tolerance. Monitoring is handled by Prometheus and Grafana.

Presentation Servers

These servers host the web application that provides access to the analyzed data and visualizations. They are responsible for serving the user interface and handling user authentication.

Server Name Role Operating System CPU RAM Storage Web Server
rainforest-web-01 Web Application Debian 11 Intel Core i7-10700K 32 GB DDR4 2 TB SSD Apache 2.4
rainforest-web-02 Web Application (Backup) Debian 11 Intel Core i7-10700K 32 GB DDR4 2 TB SSD Apache 2.4

The web application is built using Python (Flask) and JavaScript. The data is retrieved from a PostgreSQL database, documented in Database Schema. User authentication is managed through LDAP Integration.

Networking

The servers are connected via a dedicated Gigabit Ethernet network. A firewall (pfSense) protects the network from external threats. Details of the network topology can be found in the Network Diagram. We utilize a VLAN structure to segregate the different server roles.

Software Stack

The following software components are essential to the operation of the system:

⚠️ *Note: All benchmark scores are approximate and may vary based on configuration. Server availability subject to stock.* ⚠️

View full article | Back to Server rental store