What's the principle of SPARQL `SERVICE` clause query?

Question

I know that SPARQL federated query engine (such as Jena) decomposes it into individual SPARQL based on Service, then sends it to the specified SPARQL endpoint for execution, and finally combines the query results.

I know the usage of the SERVICE clause, but I don't know its principle. Do you have any relevant articles or tutorials?

For example, I have the following SPARQL federated query statement. How does it decompose and work? Can you give me a specific SPARQL sample?

The meaning of SPARQL below is: What are the Taxons under the corresponding Yago for dbo:Stephania?

SELECT ?yagoPlant
WHERE {
  SERVICE <https://dbpedia.org/sparql> {
    <http://dbpedia.org/resource/Stephania> <http://www.w3.org/2002/07/owl#sameAs> ?yago .
  }
  SERVICE <https://yago-knowledge.org/sparql/query> {
    ?yagoPlant <https://schema.org/parentTaxon> ?yago .
    ?yagoPlant a <https://schema.org/Taxon> .
  }
}
LIMIT 100