In order to query over a federation of context brokers, applications need to understand what data is contained within the brokers' collections. Then, the application can prune context brokers that are irrelevant for its query in order to optimize federated querying. GreenMov investigates how this source selection process can be achieved through the use of a Linked Data Event Stream (LDES) at the core of every context broker.
On the one hand, this LDES publishes the shape (expressed in SHACL) of every collection that is contained in a broker. On the other hand, the LDES allows third-parties to harvest the data and create an index tailored for source selection, which we call a coverage index. More specifically GreenMov investigates how the geospatial coverage of a federation of context brokers can be indexed. This is relevant for applications that request in-situ data, e.g., based on their actual location or their destination.
As iliustrated in the picture, the source selection building block uses the NGSI-LDES+Coverage Index (1) or NGSI-LD Registry API (2) to retrieve the sources that are relevant for a certain type of entity and location.
The component is under development and will be released by February 2023.