Semantic Web technologies for real-time Web of Things

The following video is a recording of a presentation, where I introduce the new support for Semantic Web technologies in the TAG Neuron®, and how it can be used to query, join and consolidate information that is available in both external and local data sources (graphs), as well as real-time communication devices, such as sensors and actuators. The Semantic Web technologies presented are standardized by the W3C, so the solution (the only solution in the world?) provides a query language for stored and real-time information in a federated network that is 100% built on standards. No proprietary languages or APIs necessary, reducing costs and complexity of implementaton.

#new, #features, #tutorial, #semantic, #web, #iot

Neurons now incorporate a SPARQL endpoint

TAG Neuron®s now incorporate a SPARQL endpoint (as of build-time 2023-07-03). SPARQL is a standardized SQL-like federated query language for information in the semantic web. SPARQL allows you to ask the entire federated network a query, which is processed in a decentralized manner, and consolidated into one response, without having to resort to collecting information into centralized repositories. With this new SPARQL endpoint, you can build systems and services that use 100% standardized interoperable queries to process information in a decentralized federated Neuron®-network. Processing information in this way represents the highest category of linked data for the web, sometimes also referred to as the five stars of open data. Combining federated web standards for the semantic web with the federated and secure nature of XMPP allows for both flexibility, interoperability and security.

Use cases

SPARQL is suitable for a wide range of applications: Consolidating information across boundaries, for instance in open data or smart city solutions, where data from different organizations need to be joined. As the semantic web standards are joined with XMPP (using the httpx URI scheme from XEP-0332), confidential and personal information can also be processed securely, without risking the data end up in the wrong hands. Queries can be processed in the secure context of digital identities of users, processors and controllers, assuring access is only granted to authorized entities.

Real-time data

The TAG Neuron® also provides access to real-time data from devices via semantic resources, making it possible to join data stored in graph data stores, with real-time data directly from the devices, without collecting and storing them in data lakes. This maximizes privacy and security, minimizes latency, and saves on communication resources and bandwith, as access to devices is only necessary when someone actually requires access, not as a means to have a relatively recent value available in a database, should someone want to access it.

Access

Access to the SPARQL endpoint is through the standardized /sparql resource on the Neuron®. Access rights to the resource needs to be granted. On a TAG Neuron® you can request access via the /Feedback.md page. You can access these on the Lab-Neuron®, via the following links:

• SPARQL Endpoint on the Lab Neuron®, also accessible via a web form.
• Request Access to the Lab Neuron®. Make sure to state full name, e-mail and purpose.
• You can also download, install and host your own Neuron®.

Graph Store (Graph Database)

The TAG Neuron® also supports the SPARQL Graph Store Protocol, which allows authorized entities to use the Neuron® to store semantic graphs on the Neuron®. They can add, update, retrieve, delete, and query such graphs. Once a graph has been added to the graph store, the URI of that graph will point to the graph stored in the graph store, and no HTTP GET to the URL will be performed, when querying the graph. This also allows for the creation of graphs with URIs that do not point to actual semantic data sources.

The resource used for accessing the graph store, is the /rdf-graph-store resource, as shown in the W3C standard specification of the graph store protocol. On the Lab Neuron®, this corresponds to https://lab.tagroot.io/rdf-graph-store. There’s a web form you can use, that interacts with the underlying graph store API also.

The default graph in the graph store consists of references to graphs in the graph store. This can be used to loop through graphs in queries, using the GRAPH pattern.

Privileges

To access the different featured mentioned, the following privileges are required to be held by the entity performing the actions:

Action	Privilege Required
SPARQL Query	Admin.Graphs.Query
Create Graph	Admin.Graphs.Add
Update Graph	Admin.Graphs.Update
Delete Graph	Admin.Graphs.Delete
Get Graph	Admin.Graphs.Get

Standardized/Interoperable Data formats

Data formats for semantic information vary. Apart from internal processing and transformation (for example for real-time data from sensors), and bespoke data formats, the following formats/representations are supported in federated queries, for interoperability:

• Semantic Information (Graphs)
  • RDF/XML
  • Turtle
  • JSON-LD
• SPARQL Query
  • SPARQL v1.2 Query format
  • Federated Queries
• SPARQL Results
  • XML
  • JSON
  • CSV
  • TSV

Data representation format is indicated using the HTTP Header Content-Type, while response type is requested using the Accept header.

Example

The following example shows a simple query for a sensor data from semantic information stored in a graph, hosted in a GitHub repository. The query does not select the graph in this example, instead it is provided via an additional input on the form. (Graphs can be specified both in the query directly, or via one or more input fields.)

Click Execute to execute the query. The result is shown dynamically below the buttons, in the format requested.

While this example is done via the web form available on the Neuron®, it can be automated via the standardized /sparql resource as well.

#neuron, #features, #api, #semantic, #web, #sparql