Microservices identification through a multi-model approach

Abstract

The life cycle of the company is characterized today by increasingly frequent phases of change induced by a continuous search for competitiveness. As a result, and in order to preserve their competitive advantages, companies must be able to react quickly to market movements, changes in the needs of their customers and business transformations of the various interlocutors. In this context, business processes playing a major role in the definition and management of information systems, must be flexible to meet the requirements of users and government regulations and all organizations must engage in the continuous maintenance of their business models. However, these models are monolithic which makes the task of maintenance difficult and very complicated Partitioning the system into microservices is usually done intuitively, based on the experience of the system designers. But, if the functionalities of a system are strongly interconnected, it is difficult to decompose the system into appropriate microservices.

To meet so this challenge, this paper proposes a multi-model based on a set of business processes. This approach combines two different independent dimensions: structural dependency and data dependency. We will be based on three clustering algorithm in order to automatically identify potential microservices.