Study: Cloud Transformation 2024

The success of cloud transformation projects is realized solely as a rise in added value for the business, which requires a precise understanding of how the cloud operating model should be strategically developed and transformed. On the one hand, there are classic variants separating IT and business even more strictly from each other; the other extreme features complete DevOps models of highly agile activities. Between these two, there are also hybrid forms with a Cloud Center of Excellence (CCoE), which acts as a kind of hub serving to link the business departments. Most importantly, a cloud transformation project must not be viewed solely from a technological perspective as the organization and working methods of IT and business departments must also adapt to the new possibilities.

The skills issue is that competencies are not rolled out across the board. There is too little understanding that virtually everyone should have cloud skills. Unfortunately, models such as a Cloud Center of Excellence often tend to move in the direction of silos. If this happens, skills are not made widely available because everyone thinks that there is a team of experts to handle any requisite tasks. Cloud skills must be very broadly defined. This begins with the classic infrastructure skills, i.e., knowledge of platform layers, requiring (for example) people who know what hyperscalers offer what database products. The data themselves, however, also demand cloud skills such as familiarity with the creation of data pipelines in Azure, AWS, or GCP. Completely new structures such as AI and machine learning are also ultimately cloud skills. And finally, frameworks such as FinOps are no less important, i.e., the cultural understanding that everyone, including developers and not just financial controllers, has to deal with costs. All these different skills need to be distributed among the right people.

I am a fan of the “start small and create value” principle. So don’t spend too much time with planning and then go straight to a major project. Realistic targets rather than overly ambitious percentages of applications that you want to have migrated by time X are key to success. An intelligent “Lift & Shift Approach” can certainly be useful here. Many people think that a fundamental transformation of the application landscape is required before the first steps in the cloud can be taken. This is not the case at the beginning, however. Experimentation, i.e., the opportunity to gain realistic experience, should also be possible. Lighthouse projects from which others can learn are also important. And the first applications should actually be running in production as opposed to having a map of the entire added-value chain in the cloud, but not having anything in live operation.

Enabling iterative approaches is crucial. Once things have been established and are poised for significant upscaling, extensive stocktaking is essential.

There is no doubt that it is becoming increasingly important. Sovereign cloud solutions are legally compliant cloud offerings that meet special requirements for data sovereignty, security, and integrity. The study reveals that about 41 percent of respondents want to ensure their “digital sovereignty.” A further 45 percent do not yet have these solutions in operation, but are already evaluating the relevant offerings on the market.

Some companies are even considering a withdrawal from the cloud for reasons of privacy and sovereignty. This is a drastic step, however, and sovereign cloud solutions should be carefully considered before taking it. Ultimately, all hyperscalers and private cloud providers offer such solutions, whereby the different aspects and variants such as privacy, confidentiality, and regulation must be evaluated from the perspective of specific requirement profiles.

The carbon footprint is certainly the most fundamental aspect. Assessments must not be limited to ongoing operations, however; the supply chain and its Scope 3 emissions must also be taken into account. Obviously, if I run a virtual machine in a data center in Sweden, my situation differs vastly from my doing the same on the East Coast of the USA with its extremely high CO₂ intensity. Yet latency and privacy considerations preclude the operation of all applications in a CO₂-friendly cloud region. But in view of GreenOps, i.e., the understanding of sustainable action, users also make an important difference because, depending on how I evolve my application, I generate higher or lower emissions. For instance, if I work more with containers or “function-as-a-service” principles compared to full virtual machines, I can also significantly reduce my carbon footprint by my own use of the cloud. In general, it should be noted that multi-cloud approaches cause additional work in the recording methodology as standard methods for assessment of the entire carbon footprint across the various clouds should be used for the company as a whole.