Improving microservices extraction using evolutionary search

doi:10.1016/j.infsof.2022.106996

Return

Improving microservices extraction using evolutionary search

2022-11-01

PRE

Khaled Sellami

Ali Ouni

Saied, Mohamed Aymen *

Salah Bouktif

Mohamed Wiem Mkaouer

DOI：10.1016/j.infsof.2022.106996

Original

Original request for help

Save

Abstract

En 中文

Context: Microservices constitute a modern style of building software applications as collections of small, cohesive, and loosely coupled services, i.e., modules, that are developed, deployed, and scaled independently. Objective: The migration from legacy systems towards the microservice-based architecture is not a trivial task. It is still manual, time-consuming, error-prone and subsequently costly. The most critical and challenging issue is the cost-effective identification of microservices boundaries that ensure adequate granularity and cohesiveness. Method: To address this problem, we introduce in this paper a novel approach, named MSExtractor, that formulates microservices identification as a multi-objective optimization problem. The proposed solution aims at decomposing a legacy application into a set of cohesive, loosely-coupled and coarse-grained services. We employ the Indicator-Based Evolutionary Algorithm (IBEA) to drive a search process towards optimal microservices identification while considering structural and semantic dependencies in the source code. Results: We conduct an empirical evaluation on a benchmark of seven software systems to assess the efficiency of our approach. Results show that MSExtractor is able to carry out an effective identification of relevant microservice candidates and outperforms three other existing approaches. Conclusion: In this paper, we show that MSExtractor is able to extract cohesive and loosely coupled services with higher performance than three other considered methods. However, we advocate that while automated microservices identification approaches are very helpful, the role of the human experts remains crucial to validate and calibrate the extracted microservices.

Keywords:

Microservices

Search-based software engineering

Legacy decomposition

Microservices architecture