Internal architecture of Microsoft Graph

Question

Microsoft Graph is very interesting.

I am interested in and curious about the internal architecture.

If you look at the use of the name Graph, it looks like you're searching for resource-to-resource(person, mail, documents..) relationships with a graph search algorithm or database, is that true?

What is the Graph API with these cases? Most of the APIs look like simple meta-search, so I want to ask you a question.

Or do you use the graph engine on the site below?

https://www.graphengine.io/