From monolithic to microservice architecture: an automated approach based on graph clustering and combinatorial optimization

Published at 2023 IEEE 20th International Conference on Software Architecture (ICSA)

Paper available at https://doi.org/10.1109/ICSA56044.2023.00013

Authors

This study has been designed, developed, and reported by the following investigators:

• Gianluca Filippone (University of L'Aquila, Italy)
• Nadeem Qaisar Mehmood (University of L'Aquila, Italy)
• Marco Autili (University of L'Aquila, Italy)
• Fabrizio Rossi (University of L'Aquila, Italy)
• Massimo Tivoli (University of L'Aquila, Italy)

For any information, interested researchers can contact us by writing an email to gianluca.filippone@graduate.univaq.it

Repository structure

Root
│   readme.md                           # This file
│   projects.json                       # Json file containing the list of the selectable monolithic systems
|   1-System_analysis.ipynb             # Jupyter Notebook implementing the static analysis step
|   2-Decomposition_optimization.ipynb  # Jupyter Notebook implementing the decomposition and optimization step
|─── javalang                           # Javalang library for java file parsing
└─── applications                       # Folder containing the considered usecases
     └─── *                             # Use case folders
          |── results                  # Folder containing the output files for the use case (results analysis included)
          └── source                   # Java source files of the use case monolithic application

Required external packages

• Javalang (https://github.com/c2nes/javalang) (included in this repository for auto-containment purposes)
• Networkx (https://networkx.org)
• Pygraphviz (https://pygraphviz.github.io)
• Gurobi (https://www.gurobi.com)
• Python-Louvain (https://github.com/taynaud/python-louvain)

Analysis step

1-System_analysis.ipynb

Input

Java code

Required user information

Refinement of the automated identification of methods, entities and system's layers

Ouptuts

• Refinement information file needed for further analysis without further refinements ([project-name]_refinement.json)
• Nodes csv file ([project-name]_nodes.csv)
• Full edges csv file, without mapping with node keys ([project-name]_edges_written.csv)
• Edges csv file, only with references with node keys ([project-name]_edges.csv)
• Graph gml file without weights ([project-name]_graph.gml)
• Graph png file ([project-name]_graph.png)

Decomposition and Optimization steps

2-Decomposition_optimization.ipynb

Input

• Graph gml file ([project-name]_graph.gml)

Optional user information

• Customization of arcs weights
• Refinement of entity clusters

Outputs

• LP formulation ([project-name]_communities_improved_optimization.lp)
• Solution csv file, mapping nodes to microservices ([project-name]_sol_communities_improved_optimization.csv)
• Solution image ([project-name]_sol_communities_improved_optimization.png)
• Solution analysis ([project-name]_sol_communities_improved_optimization.txt)

Included sources

This repository includes the Java source files of the following evaluated systems for auto-containment purposes (last update November 2022):

• Spring Petclinic (https://github.com/spring-projects/spring-petclinic)
• JPetstore (https://github.com/mybatis/jpetstore-6)
• Spring Blog (https://github.com/Raysmond/SpringBlog)