AI in Glasgow
AI in Glasgow: Server Configuration
This article details the server configuration supporting the "AI in Glasgow" project, a research initiative focused on applying artificial intelligence to urban challenges within the city of Glasgow. This document is intended for new system administrators and developers joining the project. It outlines the hardware, software, and network setup.
Overview
The "AI in Glasgow" project relies on a distributed server infrastructure to handle the demands of data ingestion, model training, and real-time inference. The system is designed for scalability and resilience, utilizing a combination of on-premise hardware and cloud resources. We utilize a hybrid approach to balance cost, security, and performance. This document will cover the core on-premise infrastructure. For details on the cloud component, please see the Cloud Integration Guide.
Hardware Configuration
The core on-premise infrastructure consists of three primary server types: Data Ingestion Servers, Training Servers, and Inference Servers. Each server type is built with specific hardware configurations optimized for its role.
| Server Type | CPU | RAM | Storage | Network Interface |
|---|---|---|---|---|
| Intel Xeon Gold 6248R (24 cores) | 128GB DDR4 ECC | 8TB RAID 10 SSD | 10GbE | ||||
| 2 x AMD EPYC 7763 (64 cores each) | 512GB DDR4 ECC | 32TB RAID 6 NVMe SSD | 100GbE | ||||
| Intel Xeon Silver 4210 (10 cores) | 64GB DDR4 ECC | 2TB NVMe SSD | 1GbE |
These servers are housed in a dedicated rack within the University of Glasgow's Data Centre. Power and cooling are managed by the data centre's infrastructure. Detailed rack diagrams are available on the Data Centre Wiki. Regular hardware maintenance schedules are outlined in the Maintenance Procedures document.
Software Stack
The software stack is built around a Linux foundation, utilizing Ubuntu Server 22.04 LTS as the operating system. Key software components include:
- Operating System: Ubuntu Server 22.04 LTS
- Containerization: Docker and Kubernetes are used for application deployment and orchestration.
- Data Storage: PostgreSQL is used for relational data, and MinIO provides object storage.
- Machine Learning Frameworks: TensorFlow, PyTorch, and scikit-learn are the primary frameworks used for model development and training.
- Monitoring: Prometheus and Grafana are used for system monitoring and alerting.
- Version Control: All code is managed using Git and hosted on GitLab.
- Networking: NGINX is used as a reverse proxy and load balancer.
Network Configuration
The server infrastructure is connected to the University of Glasgow network via a dedicated VLAN. Static IP addresses are assigned to each server. Firewall rules are configured using iptables to restrict access to necessary ports only.
| Server Role | IP Address | Subnet Mask | Gateway |
|---|---|---|---|
| 192.168.1.10 | 255.255.255.0 | 192.168.1.1 | |||
| 192.168.1.11 | 255.255.255.0 | 192.168.1.1 | |||
| 192.168.1.20 | 255.255.255.0 | 192.168.1.1 | |||
| 192.168.1.21 | 255.255.255.0 | 192.168.1.1 | |||
| 192.168.1.30 | 255.255.255.0 | 192.168.1.1 | |||
| 192.168.1.31 | 255.255.255.0 | 192.168.1.1 |
DNS resolution is handled by the University's internal DNS servers. Access to the servers from outside the University network is restricted and requires VPN access, as detailed in the Security Policy.
Security Considerations
Security is paramount. All servers are regularly patched with the latest security updates. Access to the servers is controlled via SSH keys and strong passwords. Data is encrypted both in transit and at rest. Regular security audits are conducted by the IT Security Team. Intrusion detection systems are in place to monitor for malicious activity. See the Incident Response Plan for details on handling security incidents.
Future Expansion
The infrastructure is designed to be scalable. Future expansion plans include adding more Training Servers to handle increasing model complexity and data volumes. We are also exploring the use of GPU acceleration to further improve training performance. The Capacity Planning Document outlines the projected growth and resource requirements.
| Component | Current Capacity | Projected Capacity (1 year) |
|---|---|---|
| 40TB | 80TB | ||
| 128 CPU Cores | 256 CPU Cores | ||
| 20 CPU Cores | 40 CPU Cores |
Related Documentation
- Data Pipeline Architecture
- Model Deployment Procedures
- Disaster Recovery Plan
- Backup and Restore Procedures
- Monitoring Dashboard Access
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.* ⚠️