With the exponential growth of AI technology, energy requirements are also increasing rapidly. Large technology companies are investing billions in AI accelerators quarter after quarter, leading to a surge in power consumption in data centers. In particular, the rise of generative AI and the increasing demand for graphics processing units (GPUs) have led to data centers having to scale from tens of thousands to over 100,000 accelerators.

The latest generations of AI accelerators launched by Nvidia, AMD and soon Intel have resulted in a significant increase in energy consumption per chip. For example, Nvidia's A100 has a maximum power consumption of 250W for PCIe and 400W for SXM. The successor H100 consumes up to 75 percent more, resulting in a peak power of up to 700W. This development shows that although each new generation is more powerful, it also requires more energy.

As energy consumption continues to rise with each new generation of GPUs, data centers are faced with the challenge of meeting this demand efficiently. This is where innovative cooling technologies such as liquid cooling come into play, enabling effective heat dissipation while maintaining high power density.

An important step in overcoming this challenge is the increased use of renewable energy sources. In addition, leading chip manufacturers such as Taiwan Semiconductor (TSMC) are working to improve the energy efficiency of their products. TSMC's latest manufacturing processes, such as the 3nm and the future 2nm process, promise to significantly reduce energy consumption while increasing performance.

Forecasts show that the energy requirements of AI will continue to increase in the coming years. Morgan Stanley estimates that the global energy consumption of data centers will rise to around 46 TWh in 2024, which is already a threefold increase compared to 2023. Other forecasts assume that data centers could account for up to 25 percent of total electricity consumption in the USA by 2030.

The rapid development of AI technology brings with it enormous challenges, particularly in terms of energy consumption. As a data center operator, we see it as our duty to promote and implement sustainable solutions. However, the gigantic challenges of the AI age can only be overcome together - by the united IT industry, from AI developers to chip manufacturers and data centers.

The figures speak for themselves: in the USA alone, the energy consumption of data centers could reach 35 gigawatts by 2030 - enough to power around 26 million households. Globally, AI servers could consume as much energy as Argentina, the Netherlands or Sweden by 2027.

As an industry, we must be under no illusions about the environmental costs of meeting this unprecedented demand. We need to focus all the more on sustainability - and data center operators who are not yet on board should urgently embark on the sustainable path as well.

The data center industry is no newcomer to sustainability. We have long had metrics for tracking energy and water consumption. Power Usage Effectiveness (PUE), which measures the energy consumption of a data center, was introduced in 2006. Water Usage Effectiveness (WUE) followed in 2011. These metrics encourage data centers to use resources more efficiently and at the same time offer the opportunity to save costs. Our own PUE value, for example, is 1.08, just above the ideal value of 1.0 - which means that we use almost all of our energy directly for our IT equipment and are therefore highly efficient in our use of this valuable resource.

However, reducing energy and water consumption are only two ways to achieve more sustainable data centers. In addition to operational processes, building with lower CO2 emissions can make the biggest difference - at least in the case of new facilities or upcoming renovations. In this area, we are also seeing more financing linked to sustainable outcomes.

Encouragingly, data center operators are increasingly turning to renewable energy. In the USA, over 40 gigawatts of wind and solar energy are already in use. We ourselves already operate our data center, including air conditioning, lighting and connected office space, exclusively with green electricity.

Of course, sustainable efforts not only have a positive impact on the environmental footprint of data center operators. At a time when sustainability awareness is (thankfully!) on the rise, a strong commitment to sustainability can also provide a huge competitive advantage and have a promising impact on corporate image.

At the same time, there are exciting opportunities for innovation as our industry seeks a more sustainable path. Because AI chips require so much energy to operate and cool, the technological breakthrough that reduces their energy consumption will have a multiplier effect. Even a reduction in energy consumption of just 10 percent would be a huge saving.

Ultimately, however, making data centers more sustainable is a win-win for everyone - the providers, the businesses they serve and the environment. Our industry may not have all the answers yet, but together we will surely find a better way.

Hetzner is introducing a new billing model that aims to bill products with monthly fees on an hourly basis in future. This should enable a more precise and fairer distribution of costs. However, there was one crucial detail that was missing from the original announcements: existing customers who previously had a free volume of 1,800 GB per month for cloud snapshots will now have to pay for these snapshots from May 31, 2024. According to Hetzner, this only affects a small number of customers, who were informed via a separate email.

However, Hetzner is not the first provider to charge for snapshots. Amazon Web Services (AWS) and Microsoft Azure have long since established a fee-based model for their snapshot technologies. With AWS, snapshots are stored incrementally, which means that only the changed data blocks are backed up in order to save costs. Azure uses a similar technology, with costs based on unique blocks and pages.

However, centron customers have no reason to worry. At centron, snapshots will remain free of charge in the future. At centron, snapshots are considered crucial for reliability, which is why they are made available to customers at no additional cost. After all, data security and customer satisfaction are centron's top priorities. The decision to keep snapshots free of charge underlines this commitment.

Free snapshots are a significant advantage, especially for companies that depend on reliable and free backup solutions. This enables data to be restored quickly in the event of corruption, infection or accidental deletion without incurring additional costs.

In a world where almost every application relies on digital infrastructure - from smartphone apps to critical infrastructure such as hospitals and financial services - the highly available and fail-safe operation of data centers is essential. At centron, we attach great importance to offering our customers precisely this security and reliability. Our data centers meet the highest security standards and guarantee compliance with German and European data protection laws to ensure the data sovereignty of our customers.

Another key aspect of the report is the role of data centers in promoting sustainability. Digitalization contributes significantly to the reduction of CO2 emissions by replacing analog processes and enabling more efficient solutions. At centron, we are proud to be pioneers in the use of renewable energy. Our data centers obtain the majority of their electricity from renewable sources. Across Germany, 88% of the electricity consumed by colocation data centers currently comes from renewable sources. At centron, we also rely on advanced cooling technologies to continuously improve our energy efficiency.

The Data Center Impact Report Germany 2024 further states that the data center industry creates significant economic value and contributes to Germany's digital sovereignty and economic resilience. The demand for cloud services, big data analytics and AI technologies continues to drive growth. At centron, we are continuously investing in the expansion of our infrastructure to meet this demand and make our contribution to the digital transformation. The IT capacity of colocation data centers in Germany is expected to increase from the current 1.3 GW to 3.3 GW by 2029. This is also reflected in significant investments: according to the forecast, around EUR 24 billion will be invested in the expansion of colocation capacities by 2029.

Despite the positive developments, the industry faces considerable challenges such as high electricity costs, a shortage of skilled workers and complex regulatory requirements. centron is actively committed to finding solutions, for example by promoting training and further education in the IT sector. The shortage of skilled workers is not only one of the biggest challenges in our eyes: 65% of the companies surveyed outside the Frankfurt am Main metropolitan region cited this as the biggest hurdle in the Data Center Impact Report Germany 2024.

The GDA study impressively shows how important efficient and sustainable data centers are for Germany's digital future. Data centers are increasingly being recognized as drivers of regional development. The establishment of a data center brings numerous advantages, from fiber optic connections to the creation of new jobs and the use of waste heat for municipal heat supply. Already 28% of colocation operators surveyed use their waste heat for reuse and a further 31% plan to invest in such technologies.

At Centron, we see ourselves as an integral part of this development and are working to make our data centers even more sustainable and efficient. We are actively committed to promoting a sustainable digital infrastructure.

Centron - your partner for a sustainable digital future!

The introduction of SE from the third quarter of 2025 marks an important turning point: users must have a suitable subscription license or an active Software Assurance contract as part of volume licensing. This change follows the SharePoint Server Subscription Edition model and is part of Microsoft's Modern Lifecycle Policy, which promises to continuously update the product as demand arises.

With the CU15 for version 2019, which will be released later this year, the Exchange Server will receive support for TLS 1.3 and bring back certificate management in the Exchange Admin Center (EAC). These changes will allow administrators to work more efficiently with certificates by requesting new certificates, finalizing received certificates, and exporting and importing RPX files.
It is also interesting to note that the CU15 removes support for the Unified Communications Managed API 6.0 and the instant messaging feature in the web version of Outlook, indicating the prioritization of newer technologies.

Shortly after the introduction of the SE, with CU1 in October 2025, Kerberos will be introduced as the standard protocol for server-to-server communication and will replace NTLMv2. This update will also introduce a new Admin API and remove Remote PowerShell, which was already discontinued at the end of 2022 for security reasons.

Microsoft sets out the recommended upgrade path in detail in its roadmap: Users should ideally upgrade to version 2019 CU14 with Windows Server 2022 now, before switching to Exchange Server 2019 CU15 when the new Windows Server 2025 operating system is released. The direct switch to SE then takes place from CU14 or CU15. For Exchange 2016 users, there is no direct upgrade option to SE, which requires a faster migration to the 2019 version.

The upcoming changes to Exchange Server not only mean technical updates for users, but also a significant change in licensing. The new Subscription Edition promises continuous updates and adaptations to modern technologies, but also requires a switch to the subscription model, which could pose a challenge for some organizations. The transition should therefore be planned well in advance to ensure a seamless transition.

Source: heise online

1. Using the HTTP Methods:
GET to read data.
POST to create resources.
PUT to update existing resources.
DELETE to delete resources.
Other methods such as PATCH, OPTIONS and HEAD should be used according to their specific use cases.

2. Descriptive URIs:
URIs (Uniform Resource Identifiers) should be descriptive and represent resources, not actions. Example: `/users` for user resources or `/products` for product resources.

3. Naming Resources with Nouns:
Plural nouns are standard, i.e. `/users`, `/products`.

4. Introduce Versioning:
By inserting the API version in the URI, e.g. `/api/v1/users`, changes can be implemented without affecting existing clients.

Using HTTP status codes correctly: Suitable codes such as 200 OK, 201 Created, and 500 Internal Server Error signal the result of an API operation.

JSON as a data exchange format: JSON is lightweight, easy to parse and widely used.

Use HTTP headers: These are used to transfer metadata and control caching, authentication and content type.

Standardized response format: Consistent structures for success and error responses facilitate parsing by clients.

Pagination for large data sets: Pagination should be implemented to improve performance and reduce the load on the client and server.

Authentication and authorization: Mechanisms such as OAuth and JWT (JSON Web Tokens) secure the API. Authorization mechanisms regulate access based on user roles and authorizations.

Error handling: Informative error messages and appropriate HTTP status codes are essential.

Comprehensive documentation: Tools such as Swagger or Redocly support the documentation of the API, including endpoints, request/response formats and authentication mechanisms.

API testing: Thorough testing of the API in positive and negative scenarios is essential to ensure robustness.

Fast API development with low-code tools: Tools such as Linx enable rapid development thanks to ready-made specifications and drag-and-drop interfaces.

Adhering to these guidelines and using suitable tools enables the development of reliable APIs that are not only functional but also future-proof. Although technologies evolve, the basic principles of API development remain constant and form the foundation for successful software projects.

Microservices are ubiquitous in the IT industry and are often seen as the modern standard for scalable and robust software solutions. Thanks to technologies such as cloud solutions and containerization, especially through Docker and Kubernetes, microservices have gained popularity. This architecture promises better scalability and flexibility by dividing large applications into smaller, independently manageable services. However, microservices also come with challenges:
- Complex service communication and API management
- Increased resource requirements
- Difficulties with debugging
- Potentially high costs

Many companies implement microservices without fully understanding the challenges involved, which can lead to inefficient and expensive solutions.

In contrast to this is the modular monolith, an architecture that offers advantages for small to medium-sized projects in particular. A modular monolith consists of independent modules, each of which contains everything necessary for its functionality and has clearly defined interfaces. This structure offers several advantages:
- Reduced complexity compared to microservices
- Easier refactoring and maintenance
- Lower operating costs

Especially when it comes to the development of simple CRUD systems or MVPs, a modular monolith often proves to be sufficient. Modularity also makes it possible to convert individual modules into microservices if required as soon as the system and requirements grow.

The integration of Domain-Driven Design (DDD) into a modular monolith can further improve its structure. DDD focuses on developing software around the core processes and rules of a specific business area. In a modular monolith, each module can correspond to a so-called "bounded context", which clearly structures the business logic and facilitates future extensions.

Although microservices have their place in large-scale projects such as Netflix or Shopify, a modular monolith is often the better choice for most small to medium-sized companies. This architecture offers a balanced combination of modularity, simpler management and cost efficiency. The key is that each company chooses the right architecture form based on its specific requirements and resources.