A data model in this context involves semantic structures that can be used to create message profiles without having to consider the syntax. UN/CEFACT has created the Core Component Technical Specification (CCTS) to facilitate development of data models in a structured way. Data models should be harmonized across a domain and provide inter-operability between disparate companies and governments in business-to-business (B2B) and business-to-government (B2G) scenarios.
Without a data model to contain the semantics of a business process or document, analysts cannot easily create convergence. When analysing disparate systems document requirements, a means of recording each element and its relationship with other elements has to be defined and stored in such a way that they can easily be retrieved to avoid duplication and confusion.
According to PEPPOL "a data model refers to the set of data included in a business document, and to the structure and metadata according to which the data is organized". However, it can also be a hierarchical or reference data model into which the document elements are stored and harmonized. This means that each document element is contextualised and labelled so its position in the document is known, as well as its position in the hierarchy. Each Element is inherited from a base or core component, similar to the way Classes inherit from each other and ultimately a base Class. However with CCTS, the child entity is a subset created by the restriction of elements in the parent.
According to UNECE Recommendation 18, UN/CEFACT Business Process Working Group defined the International Supply Chain in simple terms of Buy, Ship and Pay. There have been many initiatives to ensure this is contained within a reference data model and all elements within this can be linked back or contextualised to one of these terms.
Two hierarchical models that provide such structures are:
Both models use the UN/CEFACT CCTS, which ensures they are inter-operable and capable of being mapped at the semantic level. Once the semantics are agreed, then it is easier to provide a document standard using a specific syntax. Therefore it is now a prerequisite to have document standards based on robust models. This is in contrast to earlier document standards such as UN EDIFACT, where each document was not part of a hierarchical data model.