A leading German IT service provider for energy and industrial customers operates and develops complex IT infrastructures for companies with the highest requirements for availability, security, and scalability. The company is able to implement both industry-specific on-premises solutions and highly available cloud architectures and supports its customers from the planning stage through to the operation of business-critical systems. The services offered require an IT basis that can be flexibly expanded, delivers consistent performance even with strong data growth, and at the same time meets the highest data protection and compliance standards.
Project period: Q3/2025
Project volume: approx. 250,000 €
Table of Contents
Project Description
The aim of the project was to set up a highly scalable cloud service environment as the first stage of expansion at the customer's main site. The platform is based entirely on AMD EPYC 9004 processors and is supported by an all-flash storage infrastructure with a high proportion of Samsung PM9A3 and PM1743 NVMe drives. Internal connectivity is optimized via a high-speed 100 GbE network architecture to provide a high-performance and stable basis for virtual machines, container orchestration, and other cloud services.
The solution is designed so that it can be rolled out at additional locations in further project phases and expanded with powerful Ceph storage nodes. This creates a future-proof foundation for the planned standardization of an OpenStack-based cloud platform.
Key Requirements
- High computing power for virtualization and cloud workloads
- Fast, low-latency access to large amounts of data
- 100 GbE network connection for maximum bandwidth
- Scalable architecture for future locations and expansions
- High reliability and uninterrupted operation
Project Implementation
The technical implementation was carried out with specially configured MUSTANG® systems servers based on Supermicro hardware:
- Compute nodes: 10× Supermicro Hyper A+ 1125HS-TNR, each equipped with 2× AMD EPYC 9534 (64 cores, 128 threads, 2.45 GHz) and 1.152 TB DDR5-4800 ECC RAM. These nodes form the backbone of the cloud environment and provide the necessary computing capacity for virtual instances and containers.
- Manager nodes: 3× Supermicro A+ 1015CS-TNR with AMD EPYC 9124 (16 cores, 32 threads), 128 GB DDR5-4800 RAM and 960 GB Samsung PM9A3 NVMe for operating systems, as well as 2× 7.68 TB Samsung PM1743 U.2 NVMe for data. These systems perform central management and orchestration tasks.
- Control nodes: 3× Supermicro A+ 1015CS-TNR with AMD EPYC 9334 (32 cores, 64 threads), 256 GB DDR5-4800 RAM, identical storage configuration to the manager nodes. They control and monitor the platform services.
- Network connection: All systems have dual-port 100 GbE QSFP28 AIOM/OCP3 modules and support PCIe 5.0 for maximum bandwidth reserves.
- Storage: Combination of Samsung PM9A3 (960 GB) as OS drives and Samsung PM1743 (7.68 TB, PCIe Gen5) for fast, fail-safe data storage. All NVMe drives come with a 5-year manufacturer's warranty, power loss protection, and AES 256-bit hardware encryption.
Result
With the new infrastructure, the customer has a powerful, expandable cloud platform that meets the highest requirements for performance, reliability, and security. The architecture enables workloads to be distributed flexibly, minimizing latency while ensuring consistently high I/O performance thanks to all-flash storage. The prepared scaling strategy ensures that additional sites and storage resources can be seamlessly integrated—an ideal basis for the planned standardization of OpenStack cloud services.