SOA, Microservices & Integration

• Author: Sébastien Mosser (email)
• Version: 2017.09
• Reviewer: Mireille Blay-Fornarino

Case study: The Tax Computation Service

We consider here a simple case study, the Tax Computation Service. The key idea is simple: creating a service-based architecture to support the computation of income taxes for a relatively small country (i.e., Norway, 5 millions inhabitants).

Personas & Stories

• Thor is a simple tax payer, working as a software engineer in Oslo, the biggest city of the country; As Thor ...
  • I want to receive an email containing my tax return form when my taxes were processed, so that I can know the amount of taxes I paid this year;
  • I want to submit my cousin's name to the system, so that I can see how many taxes he is paying this year.
• Anders is Thor's cousin. He works as a farmer living in Fredvang, a small village located inside the Lofoten archipelago; As Anders, ...
  • I want my taxes to be handled as a farmer so that I can benefit from the special tax computation associated to this status;
  • I want to receive my tax return form by snail mail, so that I can read it without using a computer;
  • I want to receive a text message indicating that my tax return was processed.
• Eva is working at Skatteetaten, the Norwegian Tax Administration. She is supervising the tax payment process at the kingdom level; As Eva, ...
  • I want to respect the Norwegian law, stating that all processing must be done in an anonymous way;
  • I want to know how many tax return forms were processed .
• Camilla is an IT engineer, also working at Skatteetaten. She is ensuring that the computation load will not kill the tax collection process. As Camilla, ...
  • I want to supervise the operational system to monitor how the tax processing is going.

Development Timeline

Phase #1: Deploying services

• Service development
  • Creating the Tax Computation System as an RPC service;
  • Creating the Anonymous Generator as a Resource service;
  • Creating the Citizen Registry as a Document service.
• Service deployment
  • Using containers to deploy services;
  • Composing containers into a global system.
  • Monitoring containers
• Service testing
  • Acceptance testing using scenarios
  • Stress testing

Phase #2: Integrating services

• Message broker
  • Using asynchronous messages to assemble services;
  • Monitoring the broker
• Legacy integration
  • Leveraging adapters to integrate legacy systems together

Technological Stack

• Service Development:
  • Application server: TomEE+
  • REST-based service stack: JAX-RS
  • SOAP-based service stack: JAX-WS
• Integration:
  • Enterprise Service Bus: Apache Service Mix (7.0.1)
  • Message Broker: Apache ActiveMQ
  • Routing: Apache Camel (2.19.2)
• Storage:
  • Database: MongoDB (3.4)
  • Java Mapping: Jongo
• Deployment:
  • Docker Community Engine (17.07.0-rc1)
  • Docker Compose (1.15.0)
• Testing:
  • Acceptance testing: Cucumber
  • Stress testing: Gatling
• Monitoring:
  • ESB: HawtIO
  • Services: cAdvisor