Questions about data centres and sovereign AI
Answers from our data centre experts on feasibility studies, modular infrastructure, Blue DC, certification and local AI. Find the right starting point for your project.
15 answers on planning, infrastructure and AI
Data centre planning and feasibility studies
How does professional data centre planning begin?
It starts with your applications and operational requirements: which IT systems need support, what disruptions are tolerable and how should capacity grow? Site assessment, load profiles and the availability concept build on these requirements. Power, cooling, space and protection systems are considered together. We review existing documents, clarify open assumptions and develop a coordinated planning framework. This allows you to compare options before committing to a technical solution.
What does a data centre feasibility study deliver?
A feasibility study examines whether, and under what conditions, your project can be delivered at the proposed site. This includes space, power supply, cooling options, building constraints and potential expansion phases. It provides a transparent basis for decisions, setting out the options assessed, assumptions, risks and next steps. The study does not replace detailed construction design. Its scope is agreed in advance to address your question, such as a new build, expansion or modernisation during ongoing operations.
Can High Knowledge also deliver the project as general contractor?
Alongside consultancy and planning, High Knowledge provides general contracting within an agreed scope of services. Clear interfaces are essential: who designs each discipline, supplies the components and handles acceptance, documentation and operations? These responsibilities are defined for your project. Depending on your starting point, we support individual phases or coordinate several services. The basis is an agreed specification that clearly brings together requirements, deliverables and responsibilities.
Modular data centres and Blue DC
What distinguishes Blue DC E, A, M and X?
Blue DC E covers compact micro data centres and IT safe solutions. Blue DC M combines modular IT and plant areas; Blue DC X addresses larger data centre buildings. Blue DC A extends the family with a local AI solution concept based on Blue DC E. Rack count alone is not decisive: power density, cooling, redundancy, location and the application determine the appropriate design. The Blue DC overview presents the models and their technical comparison.
How is the required connection capacity calculated?
Connection capacity is derived from the planned IT load profile and its supporting infrastructure. In addition to IT, cooling, power conversion and other technical loads must be considered. Reserves and expansion phases are planned deliberately and should not be confused with expected initial utilisation. For AI applications, we also consider GPU configuration, concurrent usage and thermal load. A reliable design follows from these requirements and site conditions, rather than a generic power allowance per rack.
Can a modular data centre be expanded later?
A modular design can make expansion easier when space, connections and technical interfaces have been planned for it. Whether additional IT capacity can be integrated without interruption depends on the specific system and operating concept. We therefore consider potential expansion phases during planning, including available power, cooling reserves, cable routes, accessibility and changes during ongoing operations. This turns a general growth assumption into a technically substantiated expansion plan.
Sovereign AI and Edge AI
What does sovereign AI mean within a company?
For us, sovereign AI means that you can control key aspects of your AI environment: data access, models, infrastructure, permissions and operational processes. A local installation can help keep processing closely tied to your own environment. It does not replace careful configuration or assessment of the components used. We therefore consider the complete system, from the use case and knowledge integration to updates, logging and operational responsibilities.
When is Edge AI useful compared with centralised AI?
Edge AI brings processing closer to where data is generated or decisions are needed. Typical starting points include production sites, distributed branches and applications with limited network connectivity. The choice between local, centralised and hybrid architectures depends on data volumes, response times, availability and operational effort. We compare these requirements before selecting hardware and models. Local operation should provide a demonstrable benefit for the specific task.
Are local AI systems automatically compliant with data protection or critical infrastructure requirements?
No. Hardware location alone does not determine whether data protection, security or other requirements are met. Data types, permissions, model components, external connections and operating processes also matter. We help assess these technical and organisational issues systematically and align the architecture accordingly. The legal obligations that apply to your company and how compliance is demonstrated require separate clarification. We do not provide a blanket compliance guarantee for an AI system.
Certification, SDC and energy efficiency
How do you prepare for an assessment against DIN EN 50600 or ISO/IEC 22237?
The first step is to define the assessment subject, applicable parts of the standards and intended scope. Documentation is then reviewed to establish the evidence required for the specific infrastructure. Results and deviations must be traceable. Certification applies to the stated scope and underlying criteria; it is not a general guarantee covering every possible operating scenario. We agree the preparation, assessment and presentation of results with you.
What does Sustainable Data Center (SDC) assess?
SDC is High Knowledge’s own certification system for the structured assessment of data centres. It combines 301 criteria in eleven categories, including energy and resource efficiency, technology, organisation and operations. The specific data centre is assessed against the defined programme and its evidence, with a rating from A to G. The system name alone establishes neither climate neutrality nor full compliance with all legal obligations. Statements and scope are therefore described precisely.
What is the difference between a DC Check and certification?
A DC Check provides an initial structured review of selected aspects of your existing or planned infrastructure. It can help identify open questions and areas requiring closer assessment. Certification follows a defined assessment and decision-making process with an agreed scope and evidence requirements. The two can complement each other within a project but are not equivalent. The appropriate starting point depends on whether you need initial guidance or are already seeking a specific certification outcome.
What does PUE tell us about energy efficiency?
PUE is the ratio of a data centre’s total energy use to its IT energy use. Meaningful comparisons require consideration of measurement boundaries, time periods and utilisation. A low PUE characterises infrastructure overhead; it does not automatically establish the efficiency of IT applications or overall environmental impacts. Our calculator shows how different PUE assumptions affect an operating scenario. Results are modelled estimates, not promised savings for your facility.
Working together and starting your project
What information helps at the first project meeting?
Your objective, proposed site and a brief description of the existing IT or infrastructure are useful starting points. If available, add load assumptions, availability requirements, expansion plans and your timeline. A completed specification is not required for the first discussion. Together, we clarify what information is missing and how it can be obtained. The project enquiry helps you outline your requirements. Do not enter access credentials or confidential documents in the open enquiry field.
Where can I find reference projects and the right contacts?
Our solutions overview presents selected examples from our project work and links them to the relevant services. Further references are grouped in the construction section. Use the contact section to arrange a meeting, email us or call. Briefly describe your area of interest, such as a new build, modernisation, modular infrastructure, AI or certification. This helps us prepare a focused first discussion.
More technical questions by topic
What Exactly is a Data Center?
A data center (DC) is a purpose-built building or room used to house and operate a large amount of computing and networking equipment. The main task of a data center is to store, process and distribute data and applications. This requires a secure physical environment, a constant supply of power and cooling, an efficient network infrastructure and effective management of the hardware and software components. Data centers are often operated by large corporations, government agencies, or hosting service providers to host and manage their databases, websites, and cloud-based applications. Data centers are the basic prerequisites for digitization and thus have a significant impact on economic development. In this country, the English term “data center” or “datacenter” is very often used to designate data centers.
What Types of Data Centers are There?
There are different types of data centers that can be specialized depending on the need and intended use. Some of the common types are:
- Enterprise or corporate data centers: These data centers are operated by companies and serve to support the internal IT infrastructure.
- Colocation data centers: These are data centers operated by third parties that offer server room rental services. Companies can house their own servers or hardware in these data centers and share the infrastructure with other customers.
- Cloud data centers: These data centers are operated by providers such as Alibaba Cloud, Amazon Web Services (AWS), Google Cloud Platform (GCP), IBM Cloud, Oracle Cloud and Microsoft Azure and offer cloud-based services such as storage space, infrastructure or platforms for companies.
- Hyperscale data centers: These data centers are usually very large in scale and are suitable for large companies that need to process enormous amounts of data.
- Edge data centers: These are smaller, decentralized data centers that are close to end users and are often used for things like the Internet of Things (IoT) or smart cities.
What are Climate-Neutral Data Centers?
The climate impact of a data centre can only be assessed using clearly defined boundaries, a reporting period and traceable greenhouse gas emissions data. Depending on those boundaries, this may include construction, hardware, power supply and operations.
A low PUE, electricity from renewable sources or certification alone does not demonstrate complete climate neutrality. Emissions reductions and any offsetting must be considered separately. Claims about specific projects require appropriate evidence.
During planning, we assess energy demand, cooling, heat reuse and technical requirements in particular. This allows options and their assumptions to be compared transparently.
Server Room or Data Center – What’s the Difference?
A server room is a physical room in a building that houses servers and other IT infrastructure components such as switches, routers, and firewalls. The server room serves to provide a controlled and secure environment for the IT infrastructure. This server room is usually just a single room in a company, office or school.
In contrast, a data center is a much larger and more comprehensive facility than a single server room. A data center often consists of several server rooms and offers a comprehensive range of IT services. Data centers can house hundreds or even thousands of servers and offer additional services such as cloud hosting, storage or data backup.
Where is the Largest Data Center in the World Located?
A meaningful ranking depends on what is compared: site area, building area, operational IT capacity or the planned total capacity of a campus. Announced expansion is not the same as capacity already available for use. For your project, appropriate scale, available power, network connections, availability and expandability matter more than a size record. We derive the required scope from your load profile and operating requirements.
How Many Data Centers are There in Europe?
The number depends on the definition and reporting date. Some statistics count individual sites, others buildings or facilities above a specified IT capacity. Small server rooms, colocation locations and hyperscale campuses are therefore not directly comparable. A reliable market assessment needs to state its source, reference date, region and size categories together. For site and capacity decisions, we also consider available IT capacity, energy supply, connectivity and expansion capacity that can actually be used.
Who Needs a Data Center?
Companies, municipalities, authorities, institutions and private individuals use direct and indirect services from data centers every day. In simple terms, such systems store and calculate data. So, a data center is a place where large amounts of computer and network hardware and software are operated to store, process and manage data and provide applications to users. Companies and organizations also use data centers to maintain their services and operations by hosting databases, applications, email and other communication services. However, data centers can also be used to conduct data analysis, artificial intelligence and machine learning. They are also essential for cloud computing, as cloud providers deliver their services from data centers worldwide.
What Mistakes can be Made When Operating a Data Center?
To reduce the risk of data centre operating errors from the outset, the following points should be considered carefully:
- Power outages: A power outage can lead to an outage of the entire data center. It is important to have redundancy in the power supply to avoid failures.
- Overheating: Overheating can damage hardware components, leading to data center failure. Effective cooling of the server rooms is therefore essential.
- Network issues: Network disruptions can lead to outages of applications or services running on the servers in the data center. Careful monitoring of the network is therefore important.
- Vulnerabilities: Security gaps can lead to data loss or even data theft. It is extremely important to ensure that the entire data center is well protected against cyber attacks.
- Hardware error: Hardware failures such as defective hard drives, memory, etc. can lead to downtime. It is important to have comprehensive monitoring of all servers and components to detect errors early and fix them.
To ensure a reliable and effective data center operating environment, you must focus on thorough planning, regular maintenance, monitoring, and taking action in the event of problems.
What is Meant by “Green IT” in the Context of Data Centers?
“Green IT” Green IT refers to efforts to minimize the environmental impact of IT operations, particularly data centers. This can be achieved by using energy-efficient hardware, using renewable energy sources, optimizing cooling systems and reducing overall energy consumption. The aim is to improve the CO2 balance and promote sustainable operations.
How Important is the Choice of Location for a Data Center?
Location selection is very important for several reasons, including the availability and cost of electricity, proximity to major internet hubs for low latency, the risk of natural disasters, and climatic conditions that can affect cooling costs. A well-chosen location can significantly reduce operating costs and increase overall efficiency.
What Security Measures are Essential for Data Centers?
Essential security measures include physical security measures such as access control systems, surveillance cameras and security guards. Network-based security solutions such as firewalls, intrusion detection systems (IDS) and regular security audits are also important. Data protection compliance and redundancy in power and data connections are also crucial.
Why is Redundancy Important in Data Centers?
Redundancy is crucial to ensure service availability and reliability - even in the event of hardware failures, power outages or network problems. Multiple power supply paths, cooling systems and network connections ensure continuous operation and minimize downtime.
How Does Cooling Impact Data Center Operations?
Cooling is a key factor in energy efficiency and hardware lifespan. Effective cooling systems prevent overheating and ensure optimal performance of IT devices. Innovative cooling technologies, such as free cooling and liquid cooling, can significantly reduce cooling costs.
Which Trends Influence the Development and Operation of Data Centers?
Current trends include the use of renewable energy, the move towards edge computing to bring data processing closer to the end user, increasing automation in data center operations and the use of artificial intelligence to optimize operational processes, as well as increased attention to Sustainability and energy efficiency.
What do I Have to Consider When Planning a Data Center?
Careful planning is critical to the successful operation of a data center. For example, these aspects should be considered:
- Location: The location of the data center is important. It should be in a safe, non-seismic zone that is also easily accessible for IT staff and maintenance personnel. In addition, the local energy mix and its CO₂ emissions as well as the possible connection to local and district heating networks must be taken into account.
- Air conditioning: Since IT components can only be operated efficiently within a certain temperature range, the room climate should be continuously monitored and controlled. The data center needs a sophisticated cooling system to ensure constant temperature and humidity, as well as to set up power storage and redundancy.
- Power supply: A data center requires an uninterruptible power supply (UPS) and backup generators. A power failure could quickly lead to data loss and IT complications.
- Security: A data center must be protected against vandalism, theft or other threats through physical barriers and access systems. Access to critical areas such as the generator room, cooling system controls, etc. should be restricted to the necessary minimum.
- Network: A data center requires a fast and secure network architecture that is optimized for all systems that need to be connected in the data center.
- Monitoring: Through the use of sensors and camera surveillance systems, the data center can be monitored and secured around the clock.
- Maintenance and repair: The data center must be checked regularly for functionality and updates in order to avoid failures and data loss.
These are just a few important aspects that should be considered when planning a data center.
How is Energy Efficiency Measured in Data Centers?
Energy efficiency in data centers is often measured by the Power Usage Effectiveness (PUE) value. The PUE value is the ratio of the total energy consumed in the data center to the energy consumed by the IT devices. An ideal PUE value is close to 1.0, meaning almost no energy is spent on cooling or other non-IT-related functions.
How can the Scalability of Data Centers be Ensured?
Scalability can be ensured through modular design and the use of standardized components. This allows capacities to be flexibly added or reduced as required. Planning should also take future technologies and possible expansions of the physical location into account.
How is Data Security Guaranteed in Data Centers?
Data security is achieved through a combination of physical and digital measures, including strict access controls, encryption, regular security audits, compliance with international security standards and the implementation of security policies and procedures to prevent data leaks and cyberattacks.
Why Does a Data Center Need to be Cooled?
In order for servers and storage systems in a data center to work smoothly around the clock, they must be constantly cooled. The higher the temperature of electronic components, the greater the risk that they can fail. In addition, higher temperatures lead to excessive energy consumption, which causes high costs and is also harmful to the environment.
Storage, servers and network equipment in a data centre require electricity and release heat. Well-designed cooling removes this heat and keeps sensitive equipment within suitable operating temperatures. Energy efficiency is particularly relevant when energy prices rise and climate targets must be met. Security and resilience must also remain in focus. Appropriate redundancy helps protect operations: a cooling-system failure can otherwise rapidly lead to a major outage.
Thanks to modern technologies such as Building Information Modeling (BIM) and Data Center Management (DCM), it is possible to plan cooling systems in a network and monitor and control them from anywhere in the world. So the future of data center cooling is smart, sustainable and secure.
Why Should a Data Center be Certified?
Certification documents the assessment of a defined subject against specified requirements. The conclusions that can be drawn depend on the certification programme, assessed scope and documented results.
Audits can help identify deviations, clarify responsibilities and track improvements. A certificate does not replace controlled operations, regular maintenance or reassessment. It is not a blanket guarantee of security, environmental performance or compliance with every legal requirement.
Before certification, the objective, scope, applicable requirements and responsible certification body are therefore defined.
How is a Data Center Protected?
Protecting a data centre starts with a risk assessment and a coordinated security concept. Depending on protection needs, this includes physical security zones, access controls, early fire detection, monitoring and defined operating procedures. Network segmentation, permissions, updates and backups complement physical security. The effectiveness and interaction of these measures must be reviewed regularly. No single measure guarantees complete protection. Data protection requirements must be assessed separately in light of the data processed, access rights and processes.