The cloud shift is a fact. But cloud and established enterprise applications do not always fit together: While the infrastructure costs are falling, the established license models result in steadily increasing costs for cloud transformation on the software side. Much to the disappointment of the company on the road to cloud transformation. What is needed are holistic approaches and cloud strategies that take into account both sides – software and infrastructure – to ensure that the cloud does not fall into the cost trap. A license-optimized IT system is much more cost-effective than an IT system that is exclusively technology-optimized. Target licensing is a tried and tested approach to ensure that the digital future does not become too costly.
“Money is not an issue” has been one of the top arguments for pushing cloud solutions. Nevertheless, the discussion about the cost of cloud platforms or infrastructures and applications is by no means dead – as the cloud has become the new platform standard. However, the discussion has changed. While cloud infrastructures were considered a cost saver in their infancy, a second look revealed a different truth: “Technical flexibility has its price; it, therefore, makes sense to use the cloud for suitable (i.e. temporary and scalable) use cases”. As the Flexera 2020 State of the Cloud Report, which is representative of a plethora of other cloud analyses, noted: The number one topic on the agendas of IT managers is the optimization of the existing use of cloud resources – open bracket: cost savings. close bracket. Nevertheless, this still seems to be up for discussion.
To be fair, the discussion about controlling costs is not limited to the cloud. And the topic of cost control is being discussed on the wrong turf. To illustrate this, we should cast our minds back one (or perhaps two) eons in IT time to the days of outsourcing.
The main driving force behind the outsourcing – in addition to benefits such as focus on core competencies, higher quality, etc. – was the reduction in costs and therefore, higher efficiency: for IT, but also for personnel. Outsourcing deals were often reduced to a cost-saving percentage, which could vary anywhere between 10 and 50 percent. In outsourcing, however, a line was often drawn between infrastructures and applications, for example at the platform level. And this separation has come back to haunt us in the age of the cloud – as it has turned out to be one of the greatest challenges of cloud transformation.
The discussion about cost reductions revolves only around infrastructures and platforms – typically leaving out the application perspective. (To this day, this paradigm is reflected in a sentence we often hear: “it is up to the customer / user to supply the licenses”). This means that customers have to bear the tricky burden of license management. And buyers, license managers and software asset managers throw their hands up in despair, as they are left to pick up the pieces at the end of a decision-making process. This is because pushing the application perspective aside has serious negative cost implications for the business.
The common digital transformation strategy for IT development with a main focus on infrastructure is best distilled in the “licensing follows platform” mantra. First step: Before using any application, the application owner picks the software which will be used. Second step: Before the migration, the appropriate operating platform is selected, taking cost aspects and agility requirements into account. The third step is reduced to the gnashing of teeth over non-negotiable software licensing, which is based on the chosen platform. Software licenses influence the total cost of ownership (TCO) in the private or public cloud much more than infrastructures.
As a rule of thumb, around 60 percent of the TCO is generated by software licenses and only 40 percent by infrastructure operations. The platform focus is poison for the TCO. Enterprise software, for example from Microsoft, IBM or Oracle, is (unlike open source software) often licensed based on the “available CPUs/cores principle”; i.e. if the database or middleware runs on two cores on a larger network, the license costs will be based on the maximum number of available cores. If the platform is powerful (for example, featuring 4 CPUs with 10 cores), the resulting TCO will increase sharply.
Let us illustrate this using the continuous operation of 200 Microsoft SQL database instances. If we use a “common” (virtualized) Azure platform, the operating cost (infrastructure and licenses) amounts to € 6 million. If, on the other hand, physical cores are used in dedicated mode, the cost drops to € 4.1 million. When using a suitable dedicated future cloud infrastructure, the TCO can be reduced even further. It is easy to imagine the effect this would have on continuous operations over several years ...
In terms of a successful cloud transformation, it is therefore to think in terms of software licenses than to plan from the platform perspective. A license-optimized IT system is more cost-effective than an IT system that is exclusively technology-optimized.
This solution involves the following steps: We analyze the current mode of operation (CMO) situation of the company with regard to the current software licensing and infrastructure. In the second step, we simulate the cost-optimized future mode of operation (FMO) platform based on the CMO data collected and calculate an overall business case as a basis for the company to make decisions, enabling cost-cutting potential to be identified. This is what we call target licensing.
This approach leads to considerable cost savings - and prevents the “cloud license shock”. Turn the strategic approach around and plan under the aegis of “platform follows licensing” – and involve your license managers and buyers in your technology decisions in a manner that is “shift-left” compliant. These considerations do not play a role in pure cloud applications involving open source tools or test scenarios. However, if you want to use established enterprise software in the long-term in virtualized environments, then there is no getting around “platform follows licensing”.
