Micro-Applications

Monolithic applications have been shunned for years now, and micro-services is one of the architectures trying to replace them.

Micro-services suffer from two central problems: 1) they easily become an unmanageable soup of individually versioned components that are hard to manage and continually version, and 2) the paradigm does not really deal with user interfaces.

Micro-GUIs try to resolve the second issue, but the technology is still in the infancy.

Instead, we propose to use micro-applications. It is individual applications that work together to achieve a single goal. Micro-applications are size-wise closer to micro-services, but are self-contained in the sense that they come with data-layer and user-interface (as applicable), reducing the pain of tracing a function point between many different micro-services and allowing internal unpublished interfaces (e.g., keeping services used solely for the GUI private), reducing problem 1, and explicitly splits up the user-interface into blocks, handling problem 2.

Micro-applications allow each application to use its own paradigms and frameworks, allowing freedom in selecting not only technologies but even suppliers, yet tie everything together using a robust single-sign on solution, presenting a single experience to the end-user.

In this talk, we present the architecture in more detail and outline some of our experiences with using the paradigm at multiple customers, as well as a driving architecture for modernization of existing applications and entire landscapes.

The talk is not technical and presents all prerequisites; it is an advantage but not a prerequisite to have felt the pain of altering a 5 years untouched monolith or having to do a wide-sweeping change in a micro-service landscape.