Master Integration Testing: Why, How, and Best Practices

What is Integration Testing?

Definition (WHAT): Integration testing, also called assembly or joint testing, builds on unit testing by assembling all modules into a subsystem or system according to design requirements and testing them.

Why: Some modules work individually but may fail when connected; issues hidden in unit tests can surface at the system level.

How (basic): Combine two already‑tested units into a component and test their interface.

Zhongtong Integration Testing Concept

Focuses on testing interactions between related business functions across teams.

WHAT: Extends testing beyond a single project group to link all relevant functions across the full chain.

WHY: Business requirements often involve multiple departments; each team may deliver a correct component, yet integration can reveal risks.

HOW: A standardized process ensures smooth cross‑team integration testing.

Goals of Integration Testing

Ensure that cross‑team functionalities are correctly combined and meet overall requirements, verifying data flows, interface parameters, and integrated functionality.

Integration Testing Process

(1) Requirement Phase

Integration testing participates from the start, reviewing requirements, architecture, and documentation to identify interactions and potential defects.

(2) Test Planning

After reviews, a schedule and test plan are created, defining scope, involved systems, timeline, and personnel.

Key elements include test scope (overall and branch flows, regression), involved systems, start and end dates.

(3) Test Case Design

Based on overall sequence diagrams, requirement documents, and architecture, test cases are crafted. Design strategies include:

Finding the longest path to cover many systems.

Centering on a system and covering its radiation range.

Grouping cases for different execution strategies.

After design, a case review checks completeness, missing scenarios, and correctness.

(4) Test Execution

When all systems enter testing, integration testers run cases, record results, and report bugs to the responsible system teams or to the product manager for coordination.

(5) Test Reporting

After execution, a report is issued containing project name, owner, total and remaining defects, analysis, summary, and case execution status.

Completion Criteria

All integration test cases have been executed and passed. Unresolved defects are documented with a release note and a remediation plan, or escalated for project change decisions.

Lessons Learned and Future Work

After 2‑3 years of practice, challenges remain: understanding the difference from functional testing, allocating time, locating integration points, and ensuring end‑to‑end testing capability. Ongoing efforts focus on automation, release coordination, early warning of risky joints, and expanding testing skills.