Tutorial 3: Enterprise Data and Service Boundaries

Unit 6 ยท Enterprise security architecture

Objectives

A service boundary should clarify data ownership, API contract, identity, authorization, validation, and failure behavior. Copying sensitive data into many services increases exposure and deletion complexity. Prefer purpose-limited views, tokenized references, and explicit retention.

service A -[scoped identity + contract]-> service B
service B owns data and policy; A receives minimum representation

Distributed transactions are difficult; use idempotency, compensating actions, durable events, and clear consistency expectations. Do not let a convenience integration bypass authorization or audit.

Exercises

  1. Draw ownership for customer data across three services.
  2. Design a scoped service-to-service token.
  3. Describe failure when a downstream service is unavailable.

Self-check

  1. Why minimize data propagation?
  2. What is idempotency?
  3. Who owns a resource policy?

Self-Check Quiz

1. What should a service contract specify?

AnswerInputs, outputs, identity, authorization expectations, errors, consistency, and security constraints.

2. What does idempotency help prevent?

AnswerDuplicate effects when a request is retried after an uncertain or timed-out response.

Homework

  1. Design secure boundaries for an HR integration.
  2. Specify data minimization, identity, audit, and failure behavior.
  3. Write a threat for duplicated or stale data.
Sample answerThe HR system remains owner of employee identity; downstream services receive only necessary attributes through a scoped identity and versioned contract. Updates are authenticated, authorized, audited, idempotent, and retried with limits. Stale data has an explicit expiry and reconciliation process.