Data Center
A data center is a structurally secured building that runs servers, storage and network equipment under controlled conditions to keep IT applications permanently available. For an ERP system, the data center is the physical place where the database, documents and application logic run – whether at a cloud provider or in-house.
A data center is a specially secured and climate-controlled building, or a partitioned part of a building, in which a large number of servers, storage systems and network components are operated under controlled conditions. Its purpose is to keep IT applications and data available, secure and high-performing around the clock. Unlike a single server under a desk, a data center bundles the necessary power, cooling, network and security infrastructure so that a single defect does not bring everything to a standstill. The English term is "data center".
For business applications such as an ERP system, the data center is the physical place where the central database with product, customer and supplier master data, open orders, stock levels and postings actually runs. Even if a company only ever accesses its ERP through the browser and never hears the word "server", there is always a data center behind it. Whether it is operated by a cloud provider, a hosting service or in the company's own server room determines responsibility, cost and the question of where the data physically resides.
At a glance
- Data center = secured building for servers, storage and networking under controlled conditions
- The core is four supplies: power, cooling, network and physical security – each designed with redundancy
- Tier levels (I–IV) and availability SLAs describe how failure-resistant the operation is
- Every cloud ERP runs in a data center – the location determines where the data resides
- For the DACH region, server location, GDPR and certifications such as ISO 27001 matter
What a data center consists of
The heart of a data center is the IT systems themselves: servers that provide computing power, storage systems for the data, and network equipment that connects everything and links it to the internet. These devices sit in standardized 19-inch cabinets (racks) arranged in rows. To keep this technology running without interruption, it is surrounded by an entire floor of supply infrastructure that stays invisible in day-to-day operation but is what really sets it apart from an ordinary office server.
Four supplies are decisive here and are deliberately built multiple times over (redundantly) so that the failure of one component is absorbed. The power supply is safeguarded by uninterruptible power supplies (UPS) and emergency diesel generators, so that a grid outage does not get through for even a second. Cooling removes the enormous waste heat of the servers, because overheating would otherwise force a shutdown within seconds of a failure. The network connection runs over several independent lines from different providers. And physical security – access control, video surveillance, early fire detection and extinguishing systems – protects the facility against break-in, fire and sabotage.
Redundancy and tier levels
How failure-resistant a data center is is often described via the tier levels I to IV defined by the Uptime Institute. Tier I offers only a single supply path without redundancy, whereas Tier IV is fully fault-tolerant: every component – from the power line to the cooling – exists in duplicate, so that even maintenance or a defect during operation causes no outage. The higher the level, the lower the calculated downtime per year, but also the higher the construction cost. For ERP operators, the tier level is an indicator but not a contract – the promised availability only becomes binding in the SLA.
How a data center ensures availability
The central value of a data center is availability – the share of time during which the operated systems are actually reachable. It is expressed as a percentage and contractually guaranteed in an availability SLA. The difference sounds small but carries weight: 99.9 percent means around eight and a half hours of possible downtime per year, while 99.99 percent means just under 53 minutes. This reliability is achieved not through individual particularly good devices, but through consistent redundancy and the ability to bridge faults before they reach the user.
So that an outage at one location does not lead to total loss, large operators distribute their infrastructure across several geographically separate data centers, referred to as "availability zones" or "regions". If one location fails completely – for example due to fire or a power outage in the surrounding area – the others take over. This distribution is the technical basis for disaster recovery concepts and the reason why professional cloud services can withstand a single disaster that a single server room would not survive. A cleanly tested backup nevertheless remains mandatory, because redundancy protects against hardware failure, not against deleted or incorrect data.
The data center in the ERP system
Every ERP system runs in a data center – the only question is whose. With a cloud ERP in the SaaS model, the provider operates the application in its own or a rented data center; the company accesses it through the browser and does not have to worry about power, cooling, hardware or its maintenance. Responsibility, availability and data backup are then the provider's job and governed by the contract. With on-premise operation, by contrast, the ERP sits in the company's own server room or in a self-rented data center rack space, and the company bears full responsibility – from the air conditioning to the restore test.
When selecting an ERP system, the data center is therefore not a technical detail but a decision criterion. Where do the servers physically stand – in Germany, in the EU or outside it? What availability does the provider guarantee, and how often are backups made? Which certifications can the operator demonstrate? These points belong in the requirements specification, because they determine compliance, failure risk and, in case of doubt, the ability of the entire company to function. Scalability is another advantage of modern data centers: as the order volume grows, additional computing and storage capacity can be switched on within minutes instead of procuring new hardware over months.
Data center, cloud, server and hosting – the distinction
These terms are often mixed up. A server is a single computer that provides services; a data center is the building with all the infrastructure in which many such servers stand. Hosting refers to the service of running third-party applications or websites on this infrastructure. And cloud is the operating model in which computing power is drawn flexibly and on a usage basis from a (usually distributed) data center over the network. In short: the cloud is not a place in the clouds, but always a very concrete data center – just one the user does not have to look after themselves.
DACH specifics: location, GDPR and certification
In the German-speaking region, the location of a data center plays a special role, because personal data – for example from customer and supplier master data – is subject to the GDPR. If data is processed in a data center outside the EU, additional requirements for lawful transfer apply. Many companies therefore deliberately prefer providers with a server location in Germany, Austria or the EU – not only for legal reasons, but also because shorter distances and clear responsibilities increase trust.
Certifications serve as proof of secure operation. ISO 27001 for information security is widespread; for the DACH region, the BSI's C5 criteria catalogue, which assesses the security of cloud services, is added. Such proofs do not replace your own due diligence, but they give a solid indication that a data center's access protection, emergency preparedness and data backup are organized according to recognized standards. When selecting an ERP, it is worth having these certificates and the specific location confirmed in writing.
Cost, energy and sustainability
A data center consumes a lot of energy – not only for the servers themselves, but above all for their cooling. How efficiently a location works is measured by the PUE (Power Usage Effectiveness) metric: it relates total power consumption to pure IT consumption. A value close to 1.0 means that almost all the energy flows into the actual computing power and hardly any into ancillary consumers. Modern providers lower the PUE through free cooling with outside air, waste heat recovery and energy-efficient hardware.
For companies, this feeds directly into operating costs and is increasingly becoming part of the sustainability balance. Anyone hosting an ERP in an efficient data center powered by green electricity shifts not only responsibility but often also the ecological footprint of their own IT. In Germany, moreover, the Energy Efficiency Act (Energieeffizienzgesetz) has set minimum requirements for new data centers since 2023, for example on waste heat recovery. Such requirements will, in the medium term, also influence the pricing and location choices of cloud services.
Example
Example: An online retailer chooses the server location deliberately
A mid-sized online retailer with around 500 orders per day switches from a local server in the office to a cloud ERP. During the selection, the company compares not only features and price but asks specifically about the data center: where the servers stand, what availability is guaranteed, how often backups are made and whether ISO 27001 certification is in place.
The chosen provider operates two mirrored data centers in Frankfurt with a guaranteed availability of 99.9 percent and hourly backups. When a regional power outage hits one of the two locations a few months later, the second takes over without any noticeable interruption – the shop keeps running, orders come in. Had the old office server experienced this outage, operations would have stood still for hours. The deliberate question about the data center thus paid off directly.
Frequently asked questions
Matching ERP systems
Related services
Sources
Questions about Data Center in your ERP project?
We advise vendor-neutrally – and implement it ourselves on request.