AI in the Congo Rainforest
AI in the Congo Rainforest: Server Configuration
This article details the server configuration designed to support AI-driven research initiatives within the Congo Rainforest. This setup prioritizes reliability, data throughput, and remote accessibility, given the challenging environmental conditions and limited local infrastructure. This is geared towards newcomers familiar with basic server administration, but not necessarily with specialized AI deployments.
Overview
The project aims to deploy AI models for tasks such as species identification from audio recordings, deforestation monitoring via satellite imagery analysis, and predictive modeling of disease outbreaks. The server infrastructure is designed to handle significant data ingestion, processing, and model serving. We've opted for a distributed architecture to enhance redundancy and scalability. The central server is located in a secure, climate-controlled facility outside the rainforest, while edge servers are deployed at key research stations. We use a hybrid cloud approach, leveraging on-premise hardware and cloud services for specific workloads. See Data Acquisition for more details on data sources.
Central Server Configuration
The central server acts as the primary data repository, model training hub, and management interface. It is a high-performance machine built for intensive computation.
| Component | Specification | Quantity |
|---|---|---|
| CPU | Dual Intel Xeon Gold 6338 (32 cores/64 threads per CPU) | 2 |
| RAM | 512GB DDR4 ECC Registered 3200MHz | 1 |
| Storage (OS & Applications) | 2 x 1TB NVMe PCIe Gen4 SSD (RAID 1) | 2 |
| Storage (Data) | 32 x 18TB SAS 7.2k RPM HDD (RAID 6) | 32 |
| GPU | 4 x NVIDIA A100 80GB | 4 |
| Network Interface | Dual 100GbE QSFP28 | 2 |
| Power Supply | Redundant 2000W 80+ Platinum | 2 |
The operating system is Ubuntu Server 22.04 LTS, chosen for its stability, extensive package availability, and strong community support. We utilize Docker and Kubernetes for containerization and orchestration of AI models. Access is secured via SSH with key-based authentication and a robust firewall configured using UFW. Regular backups are performed to an offsite Cloud Storage Provider. See Security Best Practices for more details. The Database System used is PostgreSQL with a 2TB allocation.
Edge Server Configuration
Edge servers are deployed at remote research stations to perform pre-processing of data and run lightweight AI models for real-time analysis. These servers need to be robust and energy-efficient.
| Component | Specification | Quantity |
|---|---|---|
| CPU | Intel Xeon E-2388G (8 cores/16 threads) | 1 |
| RAM | 64GB DDR4 ECC 3200MHz | 1 |
| Storage (OS & Applications) | 1TB NVMe PCIe Gen3 SSD | 1 |
| Storage (Temporary Data) | 4TB SATA 7200RPM HDD | 1 |
| GPU | NVIDIA RTX A2000 12GB | 1 |
| Network Interface | Dual Gigabit Ethernet | 2 |
| Power Supply | 650W 80+ Gold | 1 |
These servers run a minimal installation of Debian and utilize MQTT for communication with the central server. AI models are deployed using TensorFlow Lite for efficient inference on resource-constrained hardware. Edge servers are powered by a combination of solar and battery power, with a backup generator for emergencies. Remote Monitoring is crucial for these servers.
Network Infrastructure
Connecting the central server and edge servers requires a reliable network. We employ a hybrid approach consisting of satellite communication and point-to-point wireless links.
| Component | Specification | Notes |
|---|---|---|
| Satellite Internet | HughesNet Gen5 | Provides primary connectivity to the central server. |
| Wireless Point-to-Point | Ubiquiti airFiber X | Connects edge servers to a central relay station. |
| Network Router (Central) | Cisco ISR 4331 | Manages network traffic and security. |
| Network Switch (Central) | Cisco Catalyst 9300 Series | Provides high-speed switching. |
| UPS (Central) | APC Smart-UPS 3000VA | Ensures power continuity. |
The network is segmented using VLANs to isolate different traffic types. We use VPNs to secure communication between the central server and edge servers. Network performance is monitored using Nagios. Firewall Rules are critically important for security. See Troubleshooting Network Issues for assistance.
Software Stack
The software stack is critical for supporting the AI workflows. This includes the operating systems as mentioned, plus:
- Python 3.10: The primary programming language.
- TensorFlow 2.9: Deep learning framework.
- PyTorch 1.12: Alternative deep learning framework.
- Pandas: Data analysis library.
- Scikit-learn: Machine learning library.
- Jupyter Notebook: Interactive development environment.
- Git: Version control system.
- Prometheus: Monitoring and alerting system.
Future Considerations
Future upgrades will include exploring the use of Federated Learning to train models directly on the edge servers, reducing the need to transfer large datasets to the central server. We are also investigating the use of Edge Computing frameworks like KubeEdge to further optimize resource utilization. Scalability Planning will be crucial as the project expands.
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 configuration
Need Assistance?
- Telegram: @powervps Servers at a discounted price
⚠️ *Note: All benchmark scores are approximate and may vary based on configuration. Server availability subject to stock.* ⚠️