AI in the French Polynesian Rainforest
# AI in the French Polynesian Rainforest: Server Configuration
This article details the server configuration supporting the "AI in the French Polynesian Rainforest" project. This project utilizes artificial intelligence to analyze biodiversity data collected from remote sensors deployed throughout the rainforests of French Polynesia. This document is intended for new system administrators and developers joining the project, providing a clear overview of the infrastructure.
Project Overview
The "AI in the French Polynesian Rainforest" project aims to monitor and understand the unique ecosystem of French Polynesia using a network of low-power sensors. These sensors collect data on audio (birdsong, insect noises), temperature, humidity, and light levels. This data is transmitted via satellite link to our central server cluster for processing and analysis. The AI models employed are primarily focused on species identification through acoustic analysis and anomaly detection indicating potential environmental changes. See Data Acquisition for details on the sensor network. Environmental Monitoring provides more background on the scientific goals.
Server Architecture
The server infrastructure is a hybrid cloud setup, leveraging both on-premise hardware for low-latency processing and cloud services for scalability and data storage. The on-premise cluster handles real-time data ingestion and initial processing, while the cloud component provides long-term storage and more complex analysis. Network Diagram details the complete architecture.
On-Premise Server Cluster
The on-premise cluster consists of three primary servers: a data ingestion server, an AI processing server, and a database server. These servers are located in a secure, climate-controlled facility in Papeete.
| Server Role | Server Name | Operating System | CPU | RAM | Storage |
|---|
| Data Ingestion | 'Moana' | Ubuntu Server 22.04 LTS | Intel Xeon Gold 6248R | 128 GB | 2 x 4TB SSD (RAID 1) |
| AI Processing | 'Raiatea' | Ubuntu Server 22.04 LTS | 2 x NVIDIA A100 GPUs, Intel Xeon Gold 6338 | 256 GB | 2 x 8TB NVMe SSD (RAID 0) |
| Database | 'BoraBora' | PostgreSQL 15 on Ubuntu Server 22.04 LTS | Intel Xeon Silver 4310 | 64 GB | 16TB HDD (RAID 6) |
These servers are interconnected via a dedicated 10 Gigabit Ethernet network. Server Security outlines the security measures implemented. The AI processing server utilizes CUDA Toolkit and TensorFlow for machine learning tasks.
Cloud Infrastructure
We utilize Amazon Web Services (AWS) for long-term data storage and backup. Specifically, we leverage:
- Amazon S3: For storing raw sensor data and processed results.
- Amazon RDS (PostgreSQL): A managed PostgreSQL database for storing metadata and analysis results.
- Amazon EC2: Used for occasional large-scale batch processing tasks.
- Programming Languages: Python 3.9, R
- Data Processing Frameworks: Apache Kafka, Apache Spark
- Machine Learning Libraries: TensorFlow, PyTorch, scikit-learn
- Database: PostgreSQL 15
- Web Server: Nginx
- Monitoring: Prometheus, Grafana (See Server Monitoring)
- Telegram: @powervps Servers at a discounted price
| AWS Service | Configuration | Approximate Cost (USD/month) |
|---|---|---|
| Amazon S3 | Standard Storage, 50TB | $125 | Amazon RDS (PostgreSQL) | db.m5.large, 20TB storage | $250 | Amazon EC2 | On-demand instance, various sizes as needed | Variable (average $100) |
Cloud Backup Strategy details the disaster recovery plan.
Software Stack
The software stack is built around open-source technologies to ensure flexibility and cost-effectiveness.
Data Flow
The data flow is as follows:
1. Sensor data is transmitted via satellite to the 'Moana' server. 2. 'Moana' server validates and pre-processes the data, then publishes it to an Apache Kafka topic. 3. The 'Raiatea' server consumes data from the Kafka topic and performs real-time AI analysis. 4. Results are stored in the 'BoraBora' database and uploaded to Amazon S3. 5. Periodic batch processing jobs on Amazon EC2 perform more complex analysis on the data stored in S3. 6. Data visualizations and reports are generated using Data Visualization Tools.
| Component | Function | Technology |
|---|---|---|
| Sensor Network | Data Collection | Low-power acoustic, temperature, humidity, and light sensors |
| Data Ingestion Server ('Moana') | Data Validation, Pre-processing, Kafka Publishing | Python, Apache Kafka |
| AI Processing Server ('Raiatea') | Real-time AI Analysis | TensorFlow, PyTorch, NVIDIA GPUs |
| Database Server ('BoraBora') | Data Storage, Metadata Management | PostgreSQL |
| Cloud Storage (Amazon S3) | Long-term Data Archival | Amazon S3 |
Future Considerations
Future plans include integrating edge computing capabilities into the sensor network to reduce latency and bandwidth requirements. We are also exploring the use of federated learning to train AI models directly on the sensor data without transmitting sensitive information to the central server. Future Development Roadmap details these plans. Please refer to Troubleshooting Guide for common issues and resolutions.
Intel-Based Server Configurations
| Configuration | Specifications | Benchmark |
|---|---|---|
| Core i7-6700K/7700 Server | 64 GB DDR4, NVMe SSD 2 x 512 GB | CPU Benchmark: 8046 |
| Core i7-8700 Server | 64 GB DDR4, NVMe SSD 2x1 TB | CPU Benchmark: 13124 |
| Core i9-9900K Server | 128 GB DDR4, NVMe SSD 2 x 1 TB | CPU Benchmark: 49969 |
| Core i9-13900 Server (64GB) | 64 GB RAM, 2x2 TB NVMe SSD | |
| Core i9-13900 Server (128GB) | 128 GB RAM, 2x2 TB NVMe SSD | |
| Core i5-13500 Server (64GB) | 64 GB RAM, 2x500 GB NVMe SSD | |
| Core i5-13500 Server (128GB) | 128 GB RAM, 2x500 GB NVMe SSD | |
| Core i5-13500 Workstation | 64 GB DDR5 RAM, 2 NVMe SSD, NVIDIA RTX 4000 |
AMD-Based Server Configurations
| Configuration | Specifications | Benchmark |
|---|---|---|
| Ryzen 5 3600 Server | 64 GB RAM, 2x480 GB NVMe | CPU Benchmark: 17849 |
| Ryzen 7 7700 Server | 64 GB DDR5 RAM, 2x1 TB NVMe | CPU Benchmark: 35224 |
| Ryzen 9 5950X Server | 128 GB RAM, 2x4 TB NVMe | CPU Benchmark: 46045 |
| Ryzen 9 7950X Server | 128 GB DDR5 ECC, 2x2 TB NVMe | CPU Benchmark: 63561 |
| EPYC 7502P Server (128GB/1TB) | 128 GB RAM, 1 TB NVMe | CPU Benchmark: 48021 |
| EPYC 7502P Server (128GB/2TB) | 128 GB RAM, 2 TB NVMe | CPU Benchmark: 48021 |
| EPYC 7502P Server (128GB/4TB) | 128 GB RAM, 2x2 TB NVMe | CPU Benchmark: 48021 |
| EPYC 7502P Server (256GB/1TB) | 256 GB RAM, 1 TB NVMe | CPU Benchmark: 48021 |
| EPYC 7502P Server (256GB/4TB) | 256 GB RAM, 2x2 TB NVMe | CPU Benchmark: 48021 |
| EPYC 9454P Server | 256 GB RAM, 2x2 TB NVMe |
Order Your Dedicated Server
Configure and order your ideal server configurationNeed Assistance?
⚠️ *Note: All benchmark scores are approximate and may vary based on configuration. Server availability subject to stock.* ⚠️