Comparative Analysis of API Testing Approaches in Service-Oriented Architectures

Abstract

This article compares three approaches to automated application programming interface (API) testing for service-oriented and microservice architectures, where application programming interfaces serve as the principal points of interaction between independent services. The growing number of such systems underscores the need for a well-founded choice of approach, since each one supports protocols and test design styles differently. The objective is to qualitatively compare a fluent assertion syntax, a typed-client harness, and an action-based scenario engine, with a focus on test design, readability, maintainability, and scope of application. The three approaches are not interchangeable products, and the comparison is drawn between the test code each of them produces rather than between feature lists. The novelty lies in a direct comparison of reference implementations that solve one common scenario, reproduced in full in the text and evaluated against criteria tied to indicators observable in the code. The analysis shows that a flexible domain-specific syntax suits the verification of Representational State Transfer (REST) interfaces, that a typed client concentrates the contract in a single declaration, and that the multi-protocol engine covers integration scenarios and message-based interaction. The choice depends on the type of interaction and the maintenance horizon of the test base. The work will be useful for test automation engineers, quality architects, and software quality assurance instructors.