Next Generation Application Modernisation

Gartner advises organizations four principles of composable business: modularity, autonomy, orchestration and discovery. On the other hand every business is tightly coupled to their application landscape, because of more than 90% of each business function is supported or implemented by software. That leads to the need of an application landscape that can be rebuild and reshaped in order to adapt and fulfil new business requirements faster.

How can we modernise our application portfolio and the underneath system landscape in order to achieve more composability?

1. If we want to arrange and rearrange pieces of a complex system in a new way, we have to under stand the dependencies between the pieces. This can be done with a dependency graph. In a dependency graph system components are nodes and their interrelationships are edges. The combination of Graph Theory and Machine Learning provides us a huge toolbox to analyse the dependencies of our living systems landscape and suggest new ways of rearranging the components. In addition EventStorming is useful to identify natural boundaries of business capabilities, processes and bounded domain contexts.
2. Change is not equal distributed across all pieces. There are software components with low rate of change and software components implementing business capabilities with high rates of change. So we have to understand which business capability or business function has or will have a high need for change. This can be done with business capability analysis.
3. Now we can follow the Stable-Dependencies Principle (SDP) of Robert Martin (one of the founder of the Manifesto of Agile Software Development). It tells us that we should implement dependencies in the direction of stability. Software components implementing Business capabilities with high rate of expected change should not be leaves in a dependency graph. Because if there are lots of dependencies directed to the business capability with high rate of change you will not be able to change it without influence of those dependent on it. In the words of the SDP business capabilities with high rate of change should not be stable, because stable things are not easy to move and can not be adapted quickly.
4. Not all dependencies are created equal. Software architecture give as a rich toolbox for implementing dependent components as loosely coupled ones. Today Cloud Technology, API Management and Domain Driven Design and Self Contained Systems are key elements in implementing more composable system landscapes.

How can we modernise our application portfolio and the underneath system landscape in order to achieve more business value?

1. Composability of the application landscape is a necessary condition for adapting fast, but it is not the Holy Grail. We always have to understand what type technology will lead to the next competitive advantage and is generating business value. Using Machine Learning (ML) to be able to make in-moment decisions will be one of the next big things. And I expect the next generation application modernisation after the ongoing cloud transformation is finished.
2. Through enhancing software components with ML the system landscape would be able learn from data and can be retrained in accordance to business needs. That is a way of "baked-in" artificial intelligence and the next step to a data driven enterprise.

Becoming a data driven enterprise with high capability to adopt Machine Learning fast and effectively will open the next generation of application modernisation cycle.