Trawler is based on the concept of a Knowledge Graph. A Knowledge Graph stores facts about the world in a graph structure. Nodes within the graph represent things - primarily data-related things like databases, tables, software jobs but also real-world things like people and teams. Edges represent relationships - e.g. dataset A is derived from dataset B or dataset C is owned by Alice. Knowledge graphs have long been studied for use within the Semantic Web and trawler makes use of a lot of these existing ideas and technologies.
Trawler's knowledge graph is based upon two components: entities (nodes) and facets (attributes on a node or edges between two nodes). The allowed types of entities and facets are (fairly loosely) restricted by an Ontology. Trawler comes with a core ontology which can describe most important facts about datasets, their schemas, their provenance etc. In addition, you can extend the ontology with your own entity and facet types.