Monolith Splitter
summary
Tool ingests execution traces and optional git lock files and outputs class clusters; backend runs Java 21 Spring Boot 3.2 with Postgres and a TypeScript React UI. Two implemented algorithms are Monot to Micro (spatio‑temporal similarity + hierarchical clustering, optimizes direct/indirect call metrics) and FOSY (functional atoms via Jaccard, non‑deterministic, optimizes intra/interconnectivity). Tested on JPetStore and Acme with clusters like card/account/orders. Limitations include need for high‑quality traces, FOSY slow on large inputs, and lack of VCS traces for some legacy projects.
AI-written summary · about
chapters (8)
- 0:00Introduction and speakers
- 0:26Problem and motivation
- 3:43Tool overview Monolith Splitter
- 6:19Algorithms MonoToMicro and FOSY
- 9:31Testing and results
- 12:23Choosing the right approach
- 13:28Limitations and challenges
- 14:38Q&A
related talks
45mMonolith to Microservices: Modularizing Fineract 1.x
46mModel Mitosis: Stop making the wrong choice between microservices and monolith
20mSplitting the Monolith
25mAuto-Detection of Copied Code from StackOverflow
45mUsing modern .NET technologies to decompose a monolith into micro-services
33mFrom Enterprise Architecture to Software Design - Critical Thinking
46mFrom Chaos to Cohesion: Building Modular Monoliths in the Real World
48mModel Mitosis: Stop Confusing Microservices and the Monolith