AI in Dundee

1. AI in Dundee: Server Configuration

This article details the server configuration powering the "AI in Dundee" initiative, providing a technical overview for system administrators and those interested in the infrastructure supporting our research. This document assumes a basic understanding of server administration and networking concepts. We will cover hardware specifications, software stack, and key configuration details.

Overview

The "AI in Dundee" project utilizes a cluster of servers dedicated to machine learning, deep learning, and data analysis tasks. The cluster is designed for scalability and high performance, employing a combination of CPU and GPU resources. The servers are located in the University of Dundee's data centre and are interconnected via a high-speed network. This infrastructure supports a variety of projects, including natural language processing, computer vision, and predictive modelling. Please refer to the Data Centre Access Policy for physical access information.

Hardware Specifications

The cluster consists of four primary server nodes, each with similar specifications. Detailed information is provided below.

Server Node	CPU	RAM	Storage	GPU
Node 1	2 x Intel Xeon Gold 6248R @ 3.0GHz	256GB DDR4 ECC REG	2 x 4TB NVMe SSD (RAID 1)	2 x NVIDIA GeForce RTX 3090 (24GB VRAM)
Node 2	2 x Intel Xeon Gold 6248R @ 3.0GHz	256GB DDR4 ECC REG	2 x 4TB NVMe SSD (RAID 1)	2 x NVIDIA GeForce RTX 3090 (24GB VRAM)
Node 3	2 x Intel Xeon Gold 6248R @ 3.0GHz	256GB DDR4 ECC REG	2 x 4TB NVMe SSD (RAID 1)	2 x NVIDIA GeForce RTX 3090 (24GB VRAM)
Node 4	2 x Intel Xeon Gold 6248R @ 3.0GHz	256GB DDR4 ECC REG	2 x 4TB NVMe SSD (RAID 1)	2 x NVIDIA GeForce RTX 3090 (24GB VRAM)

These servers are networked using 100GbE Mellanox ConnectX-6 adapters. See the Network Topology Diagram for further details. Power consumption is monitored via Power Distribution Units (PDUs).

Software Stack

The servers run Ubuntu Server 22.04 LTS. The core software stack includes:

• Operating System: Ubuntu Server 22.04 LTS
• Containerization: Docker and Kubernetes
• Machine Learning Frameworks: TensorFlow, PyTorch, scikit-learn
• Programming Languages: Python 3.9, R 4.2
• Data Storage: Ceph (distributed file system)
• Monitoring: Prometheus and Grafana

The Ceph cluster provides a scalable and resilient storage solution for the project’s data. Configuration details for Ceph can be found in the Ceph Configuration Document. Docker and Kubernetes are used for deploying and managing machine learning applications. For information on deploying applications, consult the Kubernetes Deployment Guide.

Network Configuration

The server nodes are connected to a dedicated VLAN for security and performance. Key network settings are outlined below.

Parameter	Value
VLAN ID	100
Subnet Mask	255.255.255.0
Gateway	192.168.100.1
DNS Servers	8.8.8.8, 8.8.4.4
Network Interface	enp4s0

Firewall rules are managed using `ufw` and are configured to allow only necessary traffic. The Firewall Ruleset document details the current configuration. Access to the servers is restricted to authorized personnel via SSH with key-based authentication. See the SSH Key Management Policy.

Security Considerations

Security is paramount for the "AI in Dundee" infrastructure. The following measures are in place:

• Regular security audits
• Intrusion detection system (IDS)
• Firewall protection
• Data encryption at rest and in transit
• Multi-factor authentication for administrative access

All data is backed up regularly to an offsite location. The Backup and Recovery Plan provides detailed information on the backup process. Security incidents should be reported immediately to the IT Security Team.

Future Expansion

We plan to expand the cluster with additional servers in the coming months. These new servers will feature the latest generation of GPUs and CPUs to further enhance the cluster's performance. We are also evaluating the use of NVLink technology to improve communication between GPUs. Details of the planned expansion are documented in the Future Infrastructure Roadmap. The servers will be integrated into the existing Monitoring Dashboard. We are also considering the use of a more advanced scheduling system, like Slurm, described in the Slurm Workload Manager documentation.

Main Page Server Administration Data Storage Solutions Network Security GPU Configuration

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.* ⚠️