Data-Oriented Interface Responsibility Patterns

Abstract

Remote Application Programming Interfaces (APIs) are used in almost any distributed system today, for instance in microservicesbased systems, and are thus enablers for many digitalization efforts. API design not only impacts whether software provided as a service is easy and efficient to develop applications with, but also affects the long term evolution of the software system. In general, APIs are responsible for providing remote and controlled access to the functionality provided as services; however, APIs often are also used to expose and share information. We focus on such data-related aspects of microservice APIs in this paper. Depending on the life cycle of the information published through the API, its mutability and the endpoint role, data-oriented APIs can be designed following patterns such as Operational Data Holder, Master Data Holder, Reference Data Holder, Data Transfer Holder, and Link Lookup Resource. Known uses and examples of the patterns are drawn from public Web APIs as well as application development and integration projects we have been involved in.

From OpenAPI Fragments to API Pattern Primitives and Design Smells

“…Our results are a collection of pattern primitives, which can be (and have been) composed to build larger API structures. For instance, this becomes evident when connecting the syntactical patterns that we mined here automatically with semantic architectural patterns previously mined manually by knowledge engineers [26,27]. We merely gave the first examples of primitive composition in this paper, and only outlined these connections; in our future work, we plan to investigate these topics more thoroughly.…”

“…The collection primitives that we derived from URI structures in Section 4 correspond to the MAP endpoint pattern "Information Holder" and its specializations [27]:…”