Domain Driven Service Decomposition
Use When
- Service boundaries are unclear or causing coupling
- Teams are planning monolith-to-microservice migration
Workflow
- Identify domains, subdomains, and bounded contexts.
- Map business capabilities to ownership-aligned services.
- Define context boundaries, contracts, and data ownership.
- Validate boundaries against team cognitive load and release autonomy.
- Plan incremental extraction slices with anti-corruption layers.
Required Checks
- Service boundary avoids shared write ownership
- Cross-context communication contracts are explicit
- Team ownership model supports independent deployment
Decision Framework
- Derive boundaries from business capabilities and language, not technical layers or CRUD tables.
- One bounded context owns each aggregate's write model; reads may be replicated with explicit lag tolerance.
- Size services for team cognitive load and independent release cadence—not maximal micro-granularity.
- Use anti-corruption layers when integrating legacy or external contexts.
Common Rationalizations And Rebuttals
- "Split by database table." -> Table splits ignore domain language and create chatty cross-service joins.
- "One service per developer." -> Too-fine services multiply operational cost without autonomy gains.
- "Shared kernel is faster." -> Shared mutable models reintroduce monolith coupling; share only published contracts.
Evidence Pack
- Context map with bounded contexts, upstream/downstream relationships, and ownership
- Aggregate ownership matrix (write vs read models per context)
- Extraction slice plan with strangler milestones and anti-corruption layer notes
- Team ownership alignment confirming deploy independence per slice
Exit Criteria
- Service map is implementable and evolvable
- Decomposition roadmap includes measurable migration milestones