Internet-Draft | SDF Relations | June 2024 |
Laari | Expires 22 December 2024 | [Page] |
The Semantic Definition Format (SDF) base specification defines set of basic information elements that can be used for describing a large share of the existing data models from different ecosystems. While these data models are typically very simple, such as basic sensors definitions, more complex models, and in particular bigger systems, benefit from ability to describe additional information on how different definitions relate to each other. This document specifies an extension to SDF for describing complex relationships in class level descriptions. This specification does not consider instance-specific information.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 22 December 2024.¶
Copyright (c) 2024 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
The Semantic Definition Format (SDF) [SDF] is a format for domain experts to use in the creation and maintenance of data and interaction models in the Internet of Things. The SDF specification defines a generic data model that can be used as a meta model when converting between other data models, such as IPSO Smart Objects or Digital Twins Definition Language (DTDL) [DTDL]. SDF model defines a set of affordances, describing the interfaces for the Object. These can be mapped to corresponding affordances in other data models.¶
The base specification defines ways to represent parent-child relations between two definitions. However, sometimes there is a need to describe also more complex relations to support arbitrary connections between definitions and also referring to definitions outside of the SDF models. These could be, for example, defining possible location of a device inside a room, how a device is controlled by another device, or physical topology between devices. This enables defining more complex systems using SDF models.¶
The basic parent-child relations between SDF Objects and Things can be defined by including a definition of a child in the definition of the parent. This covers a large share of simple data models defining, e.g., simple sensors, or more complex devices containing a set of sensors. On the other hand, SDF can be used also to describe even more complex entities, such as buildings with rooms and other related objects inside a building. When we extend the SDF usage, the simple parent-child relation is often not enough, but more complex relations may be needed to describe the connections between the definitions. These relations can be for example physical (e.g., an object is inside another object), functional (e.g., an object can control another object), or semantic (e.g., an object is similar to a term defined in another ontology).¶
This document extends the base SDF specification by adding a new keyword to describe also other relations between physical or logical objects than plain parent-child relations. This new keyword is needed to describe, without loss of information, models from ecosystems that are using complex relation information in their definitions.¶
This extension enables describing relations from SDF models to various (SDF or other) definitions. For a link data type for affordances, e.g., for a link property that can be accesses and modified during runtime, the "sdfType for links" extension [I-D.bormann-asdf-sdftype-link] can be used instead.¶
NOTE: This extension is now defined based on the Relationships feature in the DTDL specification. There may be other kind of definitions for relationships in other data models that must be taken into account and this specification may need to be extended to cover also those requirements.¶
This specification uses the terminology specified in [SDF], in particular "Class Name Keyword", "Object", and "Affordance".¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
This section defines a new SDF Class Name Keyword, sdfRelation, that can be used to describe complex relations. The relationship definitions are on class-level, i.e., the sdfRelation keyword does not describe instance specific information about the relation, but describes how different models and definitions relate to each other.¶
The SDF namespace block can be used to provide CURIE prefixes for external ontologies for use with sdfRelation extension. For example, in case of SAREF (Smart Applications REFerence ontology) ontology extension for buildings [saref4bldg], we can use the following namespace definition:¶
{ "namespace": { "saref": "https://saref.etsi.org/saref4bldg/v1.1.2/" } }¶
In this section, the qualities of the sdfRelation are defined. These qualities are used to define the potential type of the connection between the definitions and to which definition the connection can be made.¶
Quality | Type | Required | Description |
---|---|---|---|
relType | string | no | What kind of relationship these definitions have |
target | string | no | Target model for the relation |
description | string | no | Description of the relationship |
label | string | no | Short text describing the relationship |
property | object | no | Additional properties for this relation / e.g. from DTDL key-value pair definitions |
$comment | string | no | Additional comments for implementors |
The relType quality describes what kind of relationship this definition has to the target definition. This can use different ontologies, such as SAREF from ETSI. The used ontology MUST be defined in the namespace block to give a short name for the ontology IRI.¶
For example the "relType" field could define the relationship to be saref:isControlledByDevice
, when the SAREF ontology is used with CURIE prefix "saref" defined in the namespace block for the full IRI https://saref.etsi.org/saref4bldg/v1.1.2/
. The defined purpose for the relation is a functional relationship between the two definitions.¶
The "target" field defines to which definition or ontology term this definition with sdfRelation has a relation to. For example, this can be #/sdfObject/room
, when the target object room
is defined in the same SDF model. This may also be left undefined, and in that case the relation may be any other object (Note: This is from DTDL (check), does it make sense in SDF context?)¶
In addition to SDF definitions, the target can be also a reference to another ontology. For example, a temperature sensor SDF definition can be augmented with information that a SAREF definition of a TemperatureSensor has similar semantics as this SDF definition.¶
{ "namespace": { "exont": "https://example.com/relationOntology", "saref": "https://saref.etsi.org/core/v3.1.1/" }, "sdfObject": { "temperature": { "description": "Example temperature object", "sdfProperty": { ... }, "sdfRelation": { "sameAs": { "relType": "exont:same-as", "target": "saref:TemperatureSensor" } } } ...¶
The description of the relationship. For SDF version 1.1, the description is a string. (For future SDF versions this description can be localizable, allowing different languages in the description.)¶
In the following example, we have a definition for first-object
which located next to second-object
:¶
{ "namespace": { "exont": "https://example.com/relationOntology" }, "sdfObject": { "first-object": { "description": "Example object", "sdfProperty": { "adjacent-node": { "type": "object", "sdfType": "link" } }, "sdfRelation": { "next-to": { "description": "This object is adjacent to the second object", "relType": "exont:next-to", "target": "#/sdfObject/second-object" } } }, "second-object": { "description": "Example object, next to the first object", "sdfProperty": { "adjacent-node": { "type": "object", "sdfType": "link" } ... }, "sdfRelation": { "next-to": { "relType": "exont:next-to", "target": "#/sdfObject/first-object" } } } }¶
This section (to be removed) discusses the mapping between SDF and DTDL qualities.¶
Quality (DTDL) | Quality (SDF) | Description | Required in DTDL |
---|---|---|---|
@type | sdfRelation | DTDL Interface (Relationship), maps to sdfRelation in SDF | yes |
@id | - | DTDL: The ID of the relationship description | no |
comment | $comment | A comment for model authors | no |
description | description | DTDL: localizable description for display | no |
displayName | label | DTDL: localizable name for display | no |
maxMultiplicity | - | max multiplicity for the target, maps to maxItems in SDF instance | no |
minMultiplicity | - | min multiplicity for the target (must be zero), maps to minItems in SDF instance | no |
name | "name of relation" | The programming name of the element | yes |
properties | to sdfProperty | A set of Properties that define Relationship-specific state | no |
target | target | An interface identifier of the target (or "any" if not specified) | no |
writable | - | A boolean value that indicates whether the Relationship is writable or not, maps to SDF instance "writable" | no |
This defines the sdfRelation itself and the name is the name of the sdfRelation entry, i.e. @type Relationship and name converts to:¶
... "sdfRelation": { "name-from-DTDL": { ... } }¶
In the example DTDL files, this is never present. This is the identifier for the relationship, no further definition in the specification. In DTDL this value is given automatically if it does not exist in the DTDL model file.¶
This can be converted to $comment and it is a comment for the implementors.¶
This maps directly to the SDF "description".¶
This converts to the "label" field in SDF.¶
These define how many instances of the relationship can exist of the target type. The sdfRelation is purely a class-level definition, but sdfType "link" defines the actual instance specific information. Thus, these fields map to maxItems and minItems in the corresponding sdfType "link" definition.¶
Relationship definition in DTDL may contain additional properties (key-value pairs) that describe additional properties for this relationship. This can be converted into sdfProperty in the same object as where the sdfRelation definion is.¶
In DTDL this is the Interface of the target, in SDF this maps to the target object of this relation.¶
The relationship itself is not defined to be writable, but this field maps to the SDF instance and to the corresponding sdfType "link" definition.¶
In SDF, the relType is giving the type of the relationship, e.g. isControlledBy. However, in DTDL, this is not directly described in the DTDL file.¶
TODO Security¶
This document has no IANA actions.¶
This appendix describes the syntax of SDF relations using CDDL [CDDL]. This is providing similar description for the sdfRelation as is defined in the [SDF] for the Semantic Definition Format.¶
sdfRelation = ( sdfRelation: { ? relType: global ? target: global ? description: text ? label: text ? property: { * text => text } ? $comment: text }) ; from SDF CDDL global = text .regexp ".*[:#].*" ; rough CURIE or JSON Pointer syntax¶
The author wants to thank Ari Keranen, Mikko Saarisalo, and Christer Holmberg for their feedback and comments.¶