Standard APIs and link prediction for the digital thread

EasyChair Preprint 896

Abstract

Addressing cross-cutting concerns such as traceability, change management, and trade-off studies is complex in multi-disciplinary engineering. These aspects can only be addressed efficiently if organizations take a global level perspective on data and view data from many different sources as a whole.

Traditionally, enterprise software applications are coupled to a specific data store. The application layer communicates with the data layer through an interface to perform basic create-read-update-delete (CRUD) operations. These operations are common to most applications. By standardizing the interface between application and data layers through a standard API, enterprise software applications can be decoupled from specific data stores. APIs in general have 2 aspects: the API protocol and the semantics of the data being consumed and produced by the API. The latter aspect is independent of technology and is based on reaching a consensus between domain experts, which can be much harder than standardizing the API protocol using common technology.

Standardizing the API protocol can help organizations access data from various sources, reuse old data with new applications, and connect data to establish a digital thread. Several technologies such as REST, Linked Data, Hypermedia APIs, GraphQL, or collections of standards such as OSLC and Solid can be used to standardize the API protocol. The main ideas of these technologies as well as their adoption levels are presented. The impact of a standard API protocol for organizations, application vendors, and standardization organizations is explained. The opportunities as well as remaining challenges related to a standard API ecosystem are described.

Keyphrases: Application Programming Interface (API), Digital Thread, data integration, distributed data

@booklet{EasyChair:896,
  author    = {Axel Reichwein and Guillermo Jenaro-Radaban and Zsolt Lattmann},
  title     = {Standard APIs and link prediction for the digital thread},
  doi       = {10.29007/qvd2},
  howpublished = {EasyChair Preprint 896},
  year      = {EasyChair, 2019}}