Skip to content

Latest commit



67 lines (45 loc) · 8.58 KB

File metadata and controls

67 lines (45 loc) · 8.58 KB
title permalink
Context Map Suggestions with Service Cutter
Note: This feature has been deactivated with release v6.10.0 of Context Mapper.

The Service Cutter tool proposes a structured way to service decomposition. It suggests how a system could be decomposed into services according to 14 prioritized coupling criteria. Domain model elements are required as input. The tool then applies graph clustering algorithms to identify possible service cuts, which are returned as output. The approach was proposed in this paper.

Context Mapper Integration

We provide a Service Cutter library, which is a fork of the original Service Cutter, to be able to offer its structured decomposition approach in Context Mapper. The library allows you to propose new decomposition suggestions in the form of CML Context Maps. The decompositions are derived from the original coupling criteria catalog.

Once you have modeled your system in CML or discovered it from existing code, you can generate new decomposition suggestions or service cuts by using the following context menu entry:

Generate New Service Cuts (Context Menu)

Input and Preconditions

Service Cutter needs the system to be described in entities and so-called nanoentities. This structure is automatically derived from your CML model. The following preconditions have to be fulfilled so that we are able to derive the basic structure of your system required by Service Cutter (ERD part):

  • Your CML model must include Bounded Contexts with entities and attributes.
    • Without attributes, we cannot derive nanoentities and Service Cutter cannot calculate decompositions.

User Representations

In addition to the basic structures (including Bounded Contexts, Aggregates, domain objects, and attributes), Service Cutter takes so-called user representations which improve the suggested service cuts immensely. The CML language offers corresponding features so that Context Mapper can derive all the user representations automatically. Of course, we can only derive individual user representations if you use the corresponding CML features. The following list shows how the representations are derived and which CML features you have to use:

Solver Configuration

Once you generated your first service cut you will find a .servicecutter.yml file in the root directory of your project: (if you work with Eclipse you have to disable the .* resources filter to see the file in the project/file/package explorer)

Service Cut Generator Configuration File

In this file you can customize the priority of each coupling criterion, as you can in Service Cutter. You can further change the clustering algorithms. Details about the config file can be found here.

Example Result

When calling Propose New Service Cut, Context Mapper will create a new CML file with a new decomposition suggestion. This is one example decomposition generated for the DDD cargo sample application:

Service Cut Generator Example Result (DDD Cargo sample application)

Besides decomposition suggestion itself, Context Mapper will generate a Graphviz DOT file (*.gv) that represents the graph that was used by Service Cutter (for the graph clustering).

Generated GraphViz File (Example Screenshot)

This file can help to understand and retrace the result produced by Service Cutter. Each edge of the graph has a comment that explains the weight value that has been calculated on the basis of the coupling criteria.

You can further use online tools such as and to illustrate the graph graphically: (just copy the file content into their editors)

Generated GraphViz File in Online Editor (Example Screenshot)

Note: The graph that Service Cutter uses gets huge very quick. In our experience these online visualization tools can often not generate the visual graphs due to their complexity and size.

Service Cutter Input File Generators

If you want to work with the original Service Cutter tool, a JHipster application, you can also use our Service Cutter generators to derive the needed input files in the JSON format from your Context Map. Simply save them as JSON files (in Context Mapper) and upload them there (in the original Service Cutter). We also describe this in our service decomposition tutorial.