Server rental store

AI in the Myanmar Rainforest

AI in the Myanmar Rainforest: Server Configuration

This document details the server configuration for the "AI in the Myanmar Rainforest" project, designed to process and analyze data collected from remote sensor networks deployed within the rainforest environment. This guide is aimed at new contributors to the project and outlines the hardware and software stack powering the AI-driven analysis.

Project Overview

The “AI in the Myanmar Rainforest” project aims to leverage artificial intelligence to monitor biodiversity, detect illegal logging, and track climate change impacts within a designated region of the Myanmar rainforest. Data from acoustic sensors, camera traps, and environmental monitors is transmitted via satellite links to our central server cluster. This cluster performs real-time data processing, model training, and anomaly detection. The processed data is then made available to researchers via a dedicated web interface. See Data Acquisition for details on the sensor network.

Server Hardware Configuration

Our server infrastructure is hosted in a secure data center with redundant power and cooling. The core components are described below.

Component Specification Quantity
CPU Intel Xeon Gold 6248R (3.0 GHz, 24 cores) 3
RAM 256GB DDR4 ECC Registered 2933MHz 3
Storage (OS/Boot) 500GB NVMe SSD 3
Storage (Data) 16TB SAS 7.2k RPM HDD (RAID 6) 12
Network Interface Dual 10 Gigabit Ethernet 3
Power Supply Redundant 80+ Platinum 1200W 3

The servers are interconnected via a dedicated 40 Gigabit Ethernet backbone. A separate Network Diagram details the network topology. We utilize a clustered file system (see Storage Configuration) to provide high availability and scalability for the large datasets.

Software Stack

The software stack is built around a Linux foundation and incorporates various open-source tools for data processing, machine learning, and web serving.

Software Version Purpose
Operating System Ubuntu Server 22.04 LTS Base OS for all servers
Database PostgreSQL 14 Data storage and management
Message Queue RabbitMQ 3.9 Asynchronous task processing
Machine Learning Framework TensorFlow 2.9 Model training and inference
Web Server Nginx 1.22 Serving the web application
Programming Language Python 3.10 Primary language for data processing and AI models

We employ Docker and Kubernetes for containerization and orchestration, enabling easy deployment and scaling of services. See Deployment Procedures for more information. The Python environment is managed with venv to ensure reproducibility.

Storage Configuration

Given the large volume of data generated by the sensor network, a robust and scalable storage solution is crucial. We employ a distributed file system built on GlusterFS.

Parameter Value Description
File System GlusterFS 9.2 Distributed file system for scalability and redundancy
Replication Factor 3 Each file is replicated across three different storage nodes.
Total Storage Capacity 192 TB Aggregate storage capacity across all nodes.
Transport Protocol TCP Communication protocol used for data transfer.
Brick Directory /data/glusterfs Location of the data bricks on each server.

Data is categorized into raw sensor data, processed data, and model outputs. See Data Management Policies for details on data retention and access control. Regular backups are performed using Bacula to ensure data durability.

Security Considerations

Security is paramount, given the sensitive nature of the data collected and the remote location of the sensors.

⚠️ *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