AI in the Taiwan Rainforest
AI in the Taiwan Rainforest: Server Configuration
This article details the server configuration deployed to support the “AI in the Taiwan Rainforest” project. This project utilizes machine learning models to analyze audio and visual data collected from remote sensors within the Taiwanese rainforest, focusing on biodiversity monitoring and early detection of illegal logging activity. This document is intended for newcomers to the server infrastructure team and outlines the hardware, software, and networking components.
Project Overview
The “AI in the Taiwan Rainforest” project requires significant computational resources for real-time data processing, model training, and data storage. We've opted for a hybrid architecture, leveraging both on-premise servers for low-latency processing of immediate sensor data and cloud-based resources for large-scale model training and archival storage. This configuration balances responsiveness with cost-effectiveness. See Data Acquisition for details on data collection methods. Further information on the AI models used can be found at Model Details.
Hardware Configuration
The on-premise servers are located within a secure, climate-controlled facility near the research station in Nantou County. The servers are designed for high availability and redundancy.
On-Premise Server Specifications
| Component | Specification | Quantity |
|---|---|---|
| CPU | Intel Xeon Gold 6338 (32 cores, 2.0 GHz) | 3 |
| RAM | 256 GB DDR4 ECC Registered | 3 |
| Storage (OS/Boot) | 512GB NVMe SSD | 3 |
| Storage (Data) | 16TB SAS HDD (RAID 6) | 6 |
| GPU | NVIDIA RTX A6000 (48GB GDDR6) | 3 |
| Network Interface | Dual 10 Gigabit Ethernet | 3 |
| Power Supply | 1600W Redundant PSU | 3 |
These servers are configured in a clustered environment using Pacemaker and Corosync to ensure high availability. Detailed information on the RAID configuration can be found at RAID Configuration.
Network Infrastructure
| Component | Specification |
|---|---|
| Core Switch | Cisco Catalyst 9500 Series |
| Edge Router | Juniper MX204 |
| Firewall | Palo Alto Networks PA-820 |
| Internal Network | 10 Gigabit Ethernet |
| External Connection | 1 Gigabit Dedicated Internet Access |
The network is segmented using VLANs to isolate the server infrastructure from the research network. See Network Segmentation for more details. The firewall configuration is documented in Firewall Rules.
Software Configuration
The operating system of choice is Ubuntu Server 22.04 LTS due to its stability, extensive package repository, and strong community support. The core software stack includes:
- Docker: For containerizing applications and ensuring consistent deployments.
- Kubernetes: For orchestrating the Docker containers and managing the cluster.
- PostgreSQL: For storing metadata and processed data.
- Python 3.10: The primary language for developing and deploying the AI models.
- TensorFlow: The machine learning framework used for model training and inference.
- Prometheus: For monitoring server performance and health.
- Grafana: For visualizing metrics collected by Prometheus.
- Telegram: @powervps Servers at a discounted price
Software Stack Details
| Software | Version | Purpose |
|---|---|---|
| Ubuntu Server | 22.04 LTS | Operating System |
| Docker | 24.0.7 | Containerization |
| Kubernetes | 1.28 | Container Orchestration |
| PostgreSQL | 15.3 | Database |
| Python | 3.10.12 | Programming Language |
| TensorFlow | 2.14.0 | Machine Learning Framework |
| Prometheus | 2.45.1 | Monitoring |
| Grafana | 9.5.2 | Visualization |
The system utilizes a CI/CD pipeline implemented with GitLab CI for automated testing and deployment. Access to the servers is secured using SSH Keys and multi-factor authentication. Detailed instructions on setting up SSH access can be found on SSH Configuration.
Data Flow
Sensor data is initially received by edge devices, pre-processed locally, and then transmitted to the on-premise servers. The on-premise servers perform real-time analysis using deployed AI models. Relevant data and alerts are then forwarded to the research team. Raw data is archived to a cloud storage solution (Amazon S3) for long-term storage and large-scale model training. See Data Archiving for more details. The cloud-based training pipeline uses AWS SageMaker.
Future Considerations
Future upgrades may include the addition of more GPU capacity and an expansion of the cloud storage solution. We are also exploring the use of federated learning to reduce the need for transferring large datasets to the cloud. See Federated Learning Research for details.
Data Acquisition Model Details Pacemaker Corosync RAID Configuration Network Segmentation Firewall Rules Ubuntu Server 22.04 LTS Docker Kubernetes PostgreSQL GitLab CI SSH Configuration Data Archiving AWS SageMaker Federated Learning Research SSH Keys
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.* ⚠️