Automatic Microservices Identification from a Set of Business Processes

Abstract

All organizations engage in ongoing maintenance of their information systems due to constant changes in users' needs and governments' regulations. However these systems are monolithic making this maintenance a nightmare. To address this monolithic nature different technologies like commercial-of-the-shelf, service-oriented architecture, and lately microservices are proposed. This paper focuses on microservices by discussing their automatic identification from a set of business processes. Thanks to business processes, control and data dependencies between their activities are extracted and then clustered together. Each cluster constitutes a candidate microservice. To illustrate and demonstrate microservice automatic identification, a case study about renting bikes in the city of Barcelona is adopted and then implemented. In term of precision, the results show how business processes as inputs permit to generate better microservices compared to other approaches discussed in the paper, as well.

A multi-model based microservices identification approach

“…This paper 6 extends the approach we discussed in [10] from three perspectives. First, we add a new semantic dependency model to the existing structural-and data-dependency models.…”

“…In fact, this is what we did and demonstrated in [10] where we designed and developed a collaborative clustering-based approach to identify microservices from a set of Business Processes (BPs). To the best of our knowledge, only Amiri [2] identifies microservices using BPs while the rest use UML class diagrams [4], legacy databases [7], log files [12], and source codes [13].…”

Expert system for automatic microservices identification using API similarity graph

“…It first extracts business processes, then finds out a read/write operation of each business process to an entity to build correlation between businesses, and finally classifies and extracts candidate microservices. Daoud et al (2020) designed a collaborative clustering-based approach to automatically identify microservices, which is also based on business processes. In 2021, Daoud et al (2021) extended their former work by considering sematic,context and fragment dependency.…”

Identification of monolith functionality refactorings for microservices migration

“…To address this problem, there is recent research on the migration of monolith to a microservices architecture, for example, [3][4][5][6][7][8][9][10] which propose techniques to identify candidate decompositions of the monolith. Additionally, they use metrics to assess the quality of the decomposition.…”