At AXPE’s API Expertise Cluster, we have defined an API Strategy aimed at establishing a set of Best Practices for API design.
As part of our article series “AXPE Reference Model for API Methodology,” we will detail each aspect of this methodology. The first article will cover Introduction and General Guidelines, followed by topics such as Design, API Error Management, and Data Modeling.
The primary goal of this methodology is to standardize and homogenize API structures to:
An API (Application Programming Interface) is an interface exposed by a system to establish a technical and functional means of interaction with multiple systems. It acts as a contract and access gateway that enables system interoperability, with the following key characteristics:
This methodology is designed for technical and functional teams with experience in API specification, particularly those using API Management tools such as Kong, Mulesoft Platform, or Azure API Management.
An OpenAPI document can be structured as a single file or divided into multiple interconnected parts using keywords like .Schema, Object, $ref, etc.
We recommend naming the specification file openapi.yaml.
At AXPE, we follow the “divide and conquer” philosophy, favoring small, modular, and highly specific structures that can be reused within a single OpenAPI document or across multiple documents.
API specifications define practices, communication protocols, and structures that establish the contract.
English is the mandatory language for defining API specifications.
This applies to URLs, objects, attributes, and documentation (descriptions, examples, etc.).
The reason is that English is the predominant language in the global IT community.
Examples should be defined at the request and response level for each operation.
They serve multiple purposes:
While not mandatory (must), we strongly recommend (should) providing clear and well-documented examples as part of the API design standard.
When custom attributes need to be defined, they must follow the format “X-“ (e.g., X-Request-ID
).
Attribute naming must be aligned with the business entity data dictionary. A well-structured API must reference a catalog of business entities and their attributes to ensure alignment with business processes, products, and services.
APIs may also contain technical or control attributes that are not part of business entities. These should be clearly identified to differentiate them from business-related information.
Each defined object must include a title attribute, with the value matching the object name.
This allows validation tools (e.g., Spectral) to recognize and enforce predefined rules.
API versioning is critical as business needs evolve.
We follow Semantic Versioning (semver.org) to manage API evolution.
Semantic versioning provides comprehensive coverage for tracking API evolution and ensuring compatibility.
At AXPE’s API Expertise Cluster, we consider these general guidelines essential for establishing the fundamental principles of our Reference Model for API Methodology.
In our next article, we will cover API design standards and best practices, including:
This version maintains the clarity and technical precision required while making it structured and easy to read. Let me know if you need adjustments for tone or emphasis!