AI in East Midlands
AI in East Midlands: Server Configuration
This article details the server infrastructure supporting the "AI in East Midlands" project, a collaborative initiative focused on deploying and researching Artificial Intelligence solutions across the region. This guide is intended for new system administrators and developers joining the project. It covers hardware specifications, software stack, networking, and security considerations.
1. Hardware Overview
The project utilizes a hybrid server architecture, combining on-premise resources with cloud-based services for scalability and redundancy. The core on-premise infrastructure is located in a secure data center in Nottingham.
The primary server cluster consists of the following:
| Server Role | Server Name | CPU | RAM | Storage |
|---|---|---|---|---|
| Compute Node 1 | em-cn01 | 2 x Intel Xeon Gold 6248R | 256 GB DDR4 ECC | 4 x 4TB NVMe SSD (RAID 10) |
| Compute Node 2 | em-cn02 | 2 x Intel Xeon Gold 6248R | 256 GB DDR4 ECC | 4 x 4TB NVMe SSD (RAID 10) |
| Database Server | em-db01 | 2 x Intel Xeon Silver 4210 | 128 GB DDR4 ECC | 2 x 8TB SAS HDD (RAID 1) |
| Network/Firewall | em-fw01 | 2 x Intel Xeon E-2224 | 32 GB DDR4 ECC | 2 x 1TB SSD (RAID 1) |
These servers are complemented by cloud instances on AWS for burst capacity and specific services like model training. See Cloud Integration for details.
2. Software Stack
The software stack is designed for deep learning and data analytics. We utilize a Linux-based operating system, with specific distributions detailed below.
| Component | Version | Purpose |
|---|---|---|
| Operating System | Ubuntu Server 22.04 LTS | Base OS for all servers |
| Containerization | Docker 24.0.5 | Application deployment and isolation |
| Orchestration | Kubernetes 1.27 | Container management and scaling |
| Deep Learning Framework | TensorFlow 2.13 | Model development and training |
| Data Science Libraries | Python 3.10, Pandas, NumPy, Scikit-learn | Data manipulation and analysis |
| Database | PostgreSQL 15 | Data storage and retrieval |
All software is managed using Ansible for automated configuration and deployment. Please refer to the Deployment Procedures wiki page for detailed instructions. The Monitoring Dashboard provides real-time system status.
3. Networking Configuration
The server network is segmented into three zones: public, DMZ, and private.
- Public Zone: Accessible from the internet, hosting only the firewall.
- DMZ: Hosts web servers and API gateways, providing controlled access to internal resources.
- Private Zone: Contains the compute nodes, database server, and other sensitive systems, accessible only from within the data center network.
| Network Zone | IP Range | Access Control |
|---|---|---|
| Public | 192.168.1.0/24 | Public internet, restricted by firewall rules. |
| DMZ | 10.0.0.0/24 | Access from Public Zone via firewall, limited access to Private Zone. |
| Private | 172.16.0.0/24 | Internal network, restricted access from DMZ and Public Zones. |
The network is managed using Cisco networking equipment. Detailed network diagrams are available on the Network Topology page. All traffic is monitored using Nagios for performance and security analysis. Please consult the Firewall Ruleset documentation before making any changes to the firewall configuration.
4. Security Considerations
Security is paramount. The following measures are in place:
- Firewall: A robust firewall protects the network from external threats.
- Intrusion Detection System (IDS): Snort monitors network traffic for malicious activity.
- Regular Security Audits: Periodic security audits are conducted to identify and address vulnerabilities.
- Data Encryption: All sensitive data is encrypted at rest and in transit.
- Access Control: Strict access control policies are enforced to limit access to sensitive systems.
All users are required to follow the Security Policy document. Any security incidents should be reported immediately to the Security Team. Regular Vulnerability Scanning is performed.
5. Future Expansion
Planned expansion includes:
- Adding GPU servers for accelerated model training (see GPU Server Procurement).
- Implementing a dedicated storage cluster for large datasets (see Storage Expansion Plan).
- Integrating with additional cloud providers for increased redundancy (see Multi-Cloud Strategy).
AI Model Deployment Data Pipelines System Administration Guide Troubleshooting Guide Contact Information
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.* ⚠️