Skip to main content

Service Cut by Dataflows

Context​

Microservices are in use or are planned to be adopted. The service cut is being planned or reconsidered.

Problem​

  • A bad service cut might lead to losing the benefits of doing microservices or to high refactoring costs later on.

Solution​

Use dataflows to drive the microservice cut.

First, identify data flows and represent them in data flow diagrams for easier reasoning about the data flows. See L62 on an example algorithm to identify microservice candidates.

Maturity​

More data needed (only 1 source).

Sources of Evidence​

L62:

  • propose a "purified dataflow-driven" composition strategy
    • semi-automated mechanism
    1. Purified Data Flow Diagram (DFD) of monolith as input
    2. Condense pDFD to decomposable DFD
    3. Identify microservice candidates
    • 2 case studies where they applied their mechanism
      • comparison to by hand: more fine-grained
        • => good basis to further adjust granularity by comibing the identified microservices
  • comparison to service cutter