How is Greenplum's partitioning different from PG's inheritance

Reading the following two:

http://www.greenplumdba.com/partitioning-in-greenplum

http://gpdb.docs.pivotal.io/4350/admin_guide/ddl/ddl-partition.html

It is clear that GP's partitioning is just a skin on top of inheritance.  Does anyone know how GP implements the insert
re-routing. Is it done via triggers, just like we do in PG. 

thanks.