CASE STUDIESNEW PLATFORM · UNIVERSITY2025

Wismar University

Around 500 machines on ten nodes. A shared platform with separate environments for three faculties.

PROJECT AT A GLANCE

Wismar University

THE TASK: Build a shared data centre for three faculties.
OUR CONTRIBUTION: Consolidate requirements, support architecture and procurement, and carry out implementation and acceptance together.
THE OUTCOME: Around 500 VMs on a shared platform. Separate faculty environments, operated by the central data centre.

THE PROJECT

Starting point
and requirements.

Wismar University’s central data centre serves three faculties. Shared infrastructure needed to coexist with separate areas of responsibility.

The central data centre needed a shared platform to avoid duplicating hardware, operations and knowledge three times. Research data, teaching systems and administration had to remain separate by faculty. The departments contributed their operational knowledge. The required architecture had to combine shared use with clear separation.

NON-NEGOTIABLE REQUIREMENTS

One shared platform. Networks, storage and permissions remain separate for each faculty.

IMPLEMENTATION

What Sysfacts
took on.

Collected requirements from all participants, from central services and faculties to operations, before technical planning began
Designed a ten-node cluster across two fire compartments. The network infrastructure shares neither nodes nor physical routes
Planned separation at three levels: dedicated network segments, storage pools and permission groups per faculty
Represented the university during procurement, formulated technical requirements and assessed bids
Built two-stage backup with Proxmox Backup Server. The first stage is close to the cluster; the second is in another building
Supported implementation and failure tests with the data centre team through final acceptance

AFTER THE PROJECT

What changed for the client.

OUTCOME AND OPERATIONS

Individual hosts and an entire fire compartment were shut down for acceptance testing. Results were documented and reviewed with the central data centre. Recovery time and data-loss objectives were defined and checked.

Around 500 virtual machines run on the platform. The data centre operates it and allocates resources to faculties. Networks, storage areas and permissions remain separate.

PROJECT DATA

virtual machines on ten nodes: 500
petabyte of raw capacity across two fire compartments: 1

Technical context

Shared data centre

The target platform serves three faculties with around 500 virtual machines. Proxmox VE and Ceph run on ten nodes across two fire compartments. Raw capacity exceeds one petabyte.

Separate environments

Each faculty has its own network segments, storage pools and permission groups. The central data centre can provide shared resources while technical responsibilities and access remain separate.

Networking and backup

Redundant network paths share neither switches nor routes. Proxmox Backup Server provides two backup stages: one near the cluster and another in a separate building.

Testing and operations

Individual hosts and an entire fire compartment were shut down for failure tests. Recovery time and data-loss objectives were defined and checked. Implementation and acceptance took place with the data centre team, which now operates the platform.