Thing Description

Task force on Thing Description

This task force is chartered to address following aspects:

Clarify and define what the TDL can be used for
Clarify and define what aspects of a Thing can be described
Agreement on the three building elements:
- What is a suitable data model underlying TDL
- How would a vocabulary used in thing descriptions look like
- What is an appropriate syntax / serialization format of TDL which is feasible for the Things (may be synchronized with the AP task force)

Currently discussed in the meetings

Following findings are based on the presentations and discussions during the web meetings:

IETF CoRE Link Format

good starting point to capture what kind of resources the REST-based server provides
“rt” attribute can be used to add/link some ‘semantics’ (e.g., as a tag or a link to an ontology)
However, there is a need of a common understanding of following aspects:
- meaning of a resource (e.g., ‘measuring temperature in Celsius in room 123’)
- meaning of the data payload in terms of data types and data structures (check SenML? or XML Schema?)
Alternative to “rt” attribute: standardize a fix resource to request the complete Thing description
- e.g., define '/meta' as resource that serves the full semantic description of the Thing

W3C RDF data model

Standardized and Universal Graph representation
Exist different kinds of serialization formats:
- JSON and JSON-LD
- XML
- N3
- Turtle
- EXI (for resource constrained devices)
Available tools (e.g., for search, reasoning etc.)
- Relevant for the resource discovery
Stream-able and can be used for semantic stream processing (see OpenIoT project)

W3C Semantic Sensor Network (SSN) Ontology

Describes sensors in terms of capabilities, measurement processes, observations and deployments;
The scope of the SSN ontology and its related use cases:
- Data discovery and linking;
- Device discovery and selection;
- Provenance and diagnosis, and
- Device operation, tasking and programming;
The SSN ontology:
- is compatible with the OGC standards, and extensible with other semantic models;
- may contribute to the TD vocabulary;
- may serve for describing Things (when Things are sensors):
  - Events (e.g., stimuli, observations);
  - Properties (e.g., capabilities, conditions, ranges);
- does not need to be wholly reused, e.g., use only observations or only sensors etc.

JSON-LD

method of transporting Linked Date using JSON
JSON-LD covers RDF
- serializiation varaint for RDF-based data
save spaces by the @context defination (comparable to prefix / namespaces usage in XML)

W3C/OGC Spatial Data on the Web working Group (SDWWG)

SDWWG is working on best practices for spatial data encoding, relation vocabularies, identity, geometry and APIs.
Planned standardization work:
- OWL-time standard - ontology for the Gregorian calendar plus temporal relationships. Based on Allen’s interval calculus.
- Semantic Sensor Networks (SSN) ontology - used widely in environmental sensing and Internet of Things.
Working Working Use Cases
The work from SDWWG (e.g., SSN, OWL-time) can be used as input for vocabularies and schemas of Thing Descriptions.

Sensor Markup Language (SenML)

What is SenML
- SenML draft defines new media types for carrying simple sensor information in a protocol such as HTTP or CoAP.
- It provides a structured and uniform mechanism to represent simple sensor measurement and device parameters.
- Device parameters including name, id, type, unit, timestamp, version...
- Software implementation using JSON or XML or EXI

Motivation
- Constrained devices could easily encode a sensor measurement into the media type (application/senml+json).
- M2M server parsing the data could relatively efficiently collect a large number of sensor measurements.

Data Model
- The data is structured as a single object (with attributes) that contains an array of entries.
- Each entry is an object that has attributes such as a unique identifier for the sensor, the time the measurement was made, and the current value.
- Serializations of the data model are defined for JSON, XML and EXI.

IPSO Smart Object

based on LWM2M Object Model
reusable Object IDs and Resource IDs
is transport protocol and encoding data format independent
basic objects represent sensors and actuators
resources are defined by an object ID, object instance ID, and resource ID
- e.g., 3303/0/5700
objects can be extended with additional properties / resources

RDF serializiation with EXI

RDF data use mainly string based values (especially for subjects and predicates)
EXI has an efficient encoding approach to handle strings and removes redundancy
- relevant standardized serialization candidate for interacting with very constrained devices (e.g., microcontrollers)
a generic RDF EXI coder can be used for arbitrary RDF data or
an application specific RDF EXI coder can be used for known RDF data structures (e.g., knwon from an ontology)
hybrid approach (generic + specific RDF EXI coder) is also possible